Special Issue: Translating 18th and 19th Century European Travel WritingNewsNewsNewsVolumesVolumesVolumesVolumesSpecial Issue: Corpora and Literary TranslationSpecial Issue: Corpora and Literary TranslationSpecial Issue: Corpora and Literary TranslationSpecial Issue: Corpora and Literary TranslationSpecial Issue: Corpora and Literary TranslationSpecial Issue: Corpora and Literary TranslationSpecial Issue: Corpora and Literary TranslationSpecial Issue: Building Bridges between Film Studies and Translation StudiesSpecial Issue: Building Bridges between Film Studies and Translation StudiesSpecial Issue: Building Bridges between Film Studies and Translation StudiesSpecial Issue: Building Bridges between Film Studies and Translation StudiesSpecial Issue: Building Bridges between Film Studies and Translation StudiesSpecial Issue: Building Bridges between Film Studies and Translation StudiesSpecial Issue: Building Bridges between Film Studies and Translation StudiesSpecial Issue: Building Bridges between Film Studies and Translation StudiesSpecial Issue: Building Bridges between Film Studies and Translation StudiesNewsNewsNewsNewsNewsNewsNewsNewsNewsNewsNewsNewsNewsNewsNewsNewsNewsPagesNewsNewsNewsNewsNewsNewsNewsReviewsEditorialsVolumesNewsNewsNewsVolumesNewsReviewsNewsSpecial Issue: The Translation of Dialects in Multimedia IIIVolumesTranslationsNewsVolumesVolumesNewsSpecial Issue: A Text of Many Colours – translating The West WingSpecial Issue: A Text of Many Colours – translating The West WingSpecial Issue: A Text of Many Colours – translating The West WingSpecial Issue: A Text of Many Colours – translating The West WingSpecial Issue: A Text of Many Colours – translating The West WingSpecial Issue: A Text of Many Colours – translating The West WingSpecial Issue: A Text of Many Colours – translating The West WingSpecial Issue: A Text of Many Colours – translating The West WingSpecial Issue: A Text of Many Colours – translating The West WingSpecial Issue: The Translation of Dialects in Multimedia IIISpecial Issue: The Translation of Dialects in Multimedia IIISpecial Issue: The Translation of Dialects in Multimedia IIISpecial Issue: The Translation of Dialects in Multimedia IIISpecial Issue: The Translation of Dialects in Multimedia IIISpecial Issue: The Translation of Dialects in Multimedia IIISpecial Issue: The Translation of Dialects in Multimedia IIISpecial Issue: The Translation of Dialects in Multimedia IIISpecial Issue: The Translation of Dialects in Multimedia IIIReviewsNewsReviewsNewsNewsNewsNewsNewsNewsReviewsNewsNewsNewsNewsNewsNews

Translating Echoes

Challenges in the Translation of the Correspondence of a British Expatriate in Beresford’s Lisbon 1815-17

By António Lopes (University of the Algarve)

Abstract & Keywords

In 1812 the Farrer family established their wool trading business in Lisbon. Samuel Farrer and, a couple of years later, James Hutchinson remained in regular correspondence with Thomas Farrer, who owned a textile mill in the vicinity of Leeds, then centre of the wool trade in England. Their correspondence, spanning the period 1812-18, offers a vivid account of life in Lisbon and its hardships and troubles in the aftermath of the Peninsular War. Those letters mirror the turbulent politics of the time and articulate an attempt to narrate otherness and the way it kept challenging their gaze. The translation of the letters has posed some challenges, especially on a stylistic level. In order to confer a sense of historical authenticity on the target-language text and to attend to the stylistic features of the source-language text, the translator has been forced to revisit the Portuguese language of the period as it was spoken and written by the urban middle class in Lisbon. In this article I discuss some of the issues, both theoretical and practical, that have arisen in the course of the translation process.

Keywords: travel writing translation, commercial correspondence, private sphere, estrangement, displacement, double disjuncture, Peninsular Wars

©inTRAlinea & António Lopes (2013).
"Translating Echoes Challenges in the Translation of the Correspondence of a British Expatriate in Beresford’s Lisbon 1815-17"
inTRAlinea Special Issue: Translating 18th and 19th Century European Travel Writing
Edited by: Susan Pickford & Alison E. Martin
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/1967

1. Introduction

The world is a book, and those who do not travel read only a page.
Saint Augustine

During my research for the British Travellers in Portugal project – an ambitious initiative that has been carried out for almost three decades by the Anglo-Portuguese Studies group at the Centre for English, Translation and Anglo-Portuguese Studies (Lisbon and Oporto) –, I chanced upon a rather curious collection of letters housed at the National Archives in Kew.[1] Written by James Hutchinson Jr. (1796 - ?), a young Yorkshire merchant working in Lisbon, and addressed to his brother-in-law, Thomas Farrer, who headed the family’s wool business back in Farnley, Leeds, these letters span a period of approximately two and a half years (from 22 July, 1815 to 29 November, 1817), at a time when Portugal was struggling hard to stand on its feet after the scale of destruction caused by the Peninsular War.

Originally, the primary purpose of my undertaking was to contribute to an anthology of translated accounts of the city of Lisbon by British travellers. This meant that a considerable portion of the original text, most of it dwelling on private affairs or matters of commerce, would have to be excised in order to leave only those passages where explicit references were made to the Portuguese capital. However, it soon became evident that the scope of the content of these letters called for a differentiated approach and so the editor commissioned me to translate the complete set. The investment in an unabridged translation would give readers the opportunity not just to satisfy their curiosity about Lisbon, but above all to gain a sense of the complexity of the historical, social and economic issues with which the letters engaged, all the more so because translation is not about impoverishing the original, but about giving it a new lease of life: translation is not just a question of making a text accessible to another community of readers by acquiring a new linguistic and cultural dimension, but above all of allowing the letters to transcend their immediacy and the original purpose for which they were written, and inscribing them in new discursive practices.

So, instead of publishing excerpts of the letters in the anthology, both the editor and I decided to publish the complete set in two issues of the Revista de Estudos Anglo-Portugueses (CETAPS, Lisbon) (see Lopes 2010). This would allow us to preserve the integrity of the letters and, given the fact that the Revista is aimed at a scholarly readership (historians, philologists, cultural anthropologists, sociologists, and so on), to invest in a more detailed and in-depth approach, marked by philological accuracy and by a consciousness of the challenges posed by the hermeneutical inquiry. This would also give me the opportunity to set my own translation agenda, not just in terms of style and method, but also in terms of the future of this project. As a matter of fact, the files contain dozens of other letters and papers written by other members or friends of the family which, in view of their historical value, are also worth translating. I decided to amass all of them with the aim of publishing the whole collection in one single volume. That work is now underway.

Since translation is necessarily always a reflexive process (in more than one sense: on the one hand, the translator has to speculate about the meanings that the source text does not immediately disclose and about the readers’ responses to his/her choices; on the other, the target text always presents itself as a mirror image of the source text), the task of rendering this piece of nineteenth-century English prose into contemporary Portuguese prompted a series of theoretical and empirical questions which I set out to explore in the present article. The next section seeks to set the letters in their political, social and economic context. The meanings they contain are rooted in a specific historical setting, which has to be revisited so as to enable the text to function simultaneously as a piece of documentary evidence and as an instance of resistance: in the case of the former, substantiating that which historiography has already validated; in the case of the latter, defying or even rebutting historical theories. The third section (‘An Englishman in Lisbon’) touches on issues of estrangement, displacement and the quest for a sense of belonging, all of which are central to travel writing. The fourth section (‘Prying into a Gentleman’s Private Correspondence’) discusses the ethics and the challenges of translating the intimacy and confidentiality of private correspondence, and how the author’s objectivity gives the translator a foothold in the factual validation of his translation. The last full section (‘Translation as a Double Disjuncture’) focuses on issues of spatiality, temporality, representation and re-representation, as well as on some of the solutions to the problems posed by the historical dimension of the texts (modes of address; anachronisms; outdated terminology; formulaic language; and the need for historical research).

2. The Letters in Context: Portugal and her British Ally 1809-20

The Farrers were one among many of the local families whose lives revolved around the woollen and worsted manufacture and trade in Yorkshire. The success of their business went hand in hand with the economic growth and technological development of the period, a process which would leave an indelible mark on the landscape of the Midlands and the North of England. These developments led to major changes in the social structure, with a generalised phenomenon of rural-urban migration meeting the industry’s need for labour (Fletcher 1919: 77-84). The Yorkshire region soon became the chief export centre for manufactured woollen goods. In a world of cut-throat competition, those who succeeded in business were of an unrelenting entrepreneurial and ambitious spirit that often looked beyond the confines of Britain.

Industrial expansion forced traders to look further afield and open up new markets; Portugal swiftly became a key destination. Since Napoleon’s Continental Blockade, decreed in 1806, was firmly in place, the first industrial nation found itself in a worrying predicament. Portugal, where Britain’s commercial stakes ran high, was also left particularly exposed. It was only through Lisbon that it was possible to gain access to the Brazilian market, which had long become the mainstay of the intensive southern Atlantic economy, responsible for the capitalisation of the European market in the Early Modern period. Besides, the Portuguese could not afford to lose the support of the old ally, whose navy provided protection for the trade routes between the metropolis and its colonies. The French invasions of Portugal pushed it to the periphery of the very empire it had founded. If the demise of both commerce and industry had a terrible impact on the economy, the destruction the war wrought in the provinces proved no less damaging. Looting, extortion and massacres left a trail of blood, hatred and revulsion across the whole nation that was to remain unabated for generations. Wellington’s scorched earth policy – aiming to deprive the French troops of victuals and other supplies – aggravated the situation even further. Agriculture and husbandry practically ground to a halt and farmers were unable to produce the foodstuffs required to feed the urban centres. Famine set in and with it a period of demographic stagnation.

Freeing Portugal from the chains of Napoleonic imperialism was not without its costs. Unable to overcome such complete vulnerability, the nation was at the mercy of British interests. Certainly a significant part of the Portuguese economy had for a long time depended on Britain. Whether Portugal benefited from this trade relationship or not is a matter of controversy (Borges de Macedo 1963; Bethell 1984; Maxwell 2004; Pijning 1997; Pardo 1992). However, at least since the Methuen Treaty (1703) Britain had been undermining the Portuguese industry with a substantial influx of cheap manufactured goods undercutting all competition. In January 1808 the opening of the Brazilian ports to Britain represented a fatal blow. Two years later, the protective mechanism of customs duties was removed precisely when the Portuguese economy was most in need of it. The prospects for the manufacturing sector grew dimmer as British cotton and wool cloths flooded the Portuguese market.

The political power that William Carr Beresford, commander-in-chief of the Portuguese troops during the invasions, held during this crucial period in the country’s history played a decisive role in protracting this position of economic subordination. He ended up gaining considerable ascendancy over the representatives of the Prince Regent. In the post-war years he headed the military government, a position which rapidly eroded his earlier prestige as a war hero. People started protesting against the way public funds were being squandered to pay for the presence of British troops on national territory. Portuguese officers likewise harboured deep-seated resentment towards the British officers, who were now apparently being granted all sorts of privileges and promotions (see Glover 1976). Beresford’s radical intransigence in politics led to the repression of those who advocated a more liberal agenda, namely those who were suspected either of sympathising with the ideals of the French Jacobins, or of defending a constitutional monarchy. As a stern defender of Tory absolutism, his views were in line with the ones shared by two other Anglo-Irish potentates, namely Wellington and Castlereagh (Newitt 2004: 107). His absolutist values, along with his thirst for power, left him isolated in a world riven by deep-rooted hatreds. The revolutionary clamour heard in Oporto on 24 August 1820 was to put paid to Beresford’s ambitions. Paradoxically, partly thanks to the influence of the British officers, the British tradition of liberalism ended up taking root in a country lacking in ideological coordinates to define its political future.

When James Hutchinson first set foot in Lisbon, the country was going through a period of economic depression. His letters mirror the upheavals and the social unrest of the period and therefore help to shed light on historical processes, since they testify to the way in which individuals perceived reality and (re)acted accordingly. Popular reactions to the new king, news of the uprising in Pernambuco (Brazil), political persecutions, and hangings are well documented elsewhere,[2] but here we are given a view from the inside. Moreover, rather than just affirming the picture that the extensive historiographical literature on the subject has already established, the letters also disclose new facets. They prove that, despite the impressive growth of Britain’s exports in this period, British trade did not run smoothly in Portugal. Hutchinson could hardly be said to be the definitive model of the successful businessman. His efforts, nonetheless, were mostly undermined by factors that lay beyond his reach. General poverty, scarcity of money, shortages of food and other essentials, and rationing, for example, became recurrent, if not obsessive, subjects in his letters, betraying his sense of frustration and underachievement. Moreover, Hutchinson was forced to deal with fierce competition within the Portuguese market and the incompetence of the Customs officials, not to mention liabilities and bad debts, marketing obstacles and, curiously enough, an increasingly demanding clientele, all of which imposed psychological costs he found ever more difficult to cope with. And although he was not so forthcoming in discussing political issues, such as Beresford’s repression, his fears and silences about the persecutions are no less telling.

Each letter contains, as it were, the very essence of history and, through the picturesque and sometimes disconcerting episodes they feature, they help us recreate a reality long buried by time. Precisely because this is a genuine voice that has remained hidden amidst other archival material for almost two centuries, unscathed by later misappropriations or misinterpretations, we are able to salvage pristine fragments of the historical experience and to retrieve for our collective memory some of the particularities and singularities that are usually overlooked in the construction of the historical grand narratives of the nation. In a letter dated 18 October 1816, for instance, Hutchinson speaks of the funeral ceremonies of Queen Maria I and clearly enjoys recounting the peculiar causes of the accidental fire that burned down the church where those ceremonies were being held. In a later letter (22 October 1817), he provides a first-hand testimony of the horrendous hanging of the men who followed Gomes Freire de Andrade in his revolt against Lord Beresford’s roughshod rule. Elsewhere he laments the shortage of foodstuffs and the rise in prices which mercilessly strike the poor (letter dated 25 January 1817), but he cannot help relishing the story of a woman arrested for stealing bodies from the cemetery to produce black pudding to be sold to the local shops (9 August 1816). In another letter he speaks of an earthquake that threw the city ‘into the most dreadful alarm’ and the scenes of panic that ensued, while rejoicing at the fact that he remained ‘during the whole of the night in a sound slumber’ (3 February 1816).

3. An Englishman in Lisbon: Estrangement, Displacement and the Quest for Belonging

Notwithstanding the rapid decline of the Portuguese economy during and after the Peninsular War, British traders rapidly resumed their investments in the country. Samuel Farrer & Sons were amongst them. Samuel Farrer Jr. established the family’s business in Lisbon in 1812. The family’s entrepreneurial effort must have paid off somehow, for upon his death, in February 1815, they decided to keep on investing in their Portuguese venture. It would be up to young James Hutchinson Jr. to take up the business. His inexperience notwithstanding, James was not entirely at a loss. The need to account for every transaction and to keep his brother-in-law posted about how business was being conducted resulted in a correspondence of considerable length, which lasted until his departure from Lisbon at the end of 1817. The letters were permeated by the young man’s comments, remarks and anecdotes about life in the Portuguese capital. Being an outsider in customs, language and feelings, Hutchinson tried hard to accommodate himself to his new setting.

In his letters, however, the affectionate attachment he exhibits towards his sister and the other members of his family indicates that his stay in Lisbon was, emotionally speaking, hard to bear. He often complained about her silence and the fact that she now seemed to have forsaken him altogether. But then, it was not just the separation from his loved ones that threw him into a state of melancholy. His life in the Portuguese capital was infused with a sense of estrangement he was unable to overcome. He felt uprooted and disengaged.

It becomes all too apparent that his gaze is that of an outsider, of someone struggling to succeed in a strange, disturbing world, whose social and political environment contrasts in many respects with that of his native land. He soon realised it would not be easy to fit in. Despite the support that other British expatriates residing in Lisbon gave him, he complained to his family about living conditions there. Blatantly ironic, he confessed that he ‘suffer[ed] very much from the Muschetos [sic], Bugs & other filth with which this sweet City so much abounds’ (11 August 1815).

His difficulty in understanding the Portuguese is particularly visible when he is faced with the lack of patriotic fervour of the man in the street, a fervour one should expect from a nation that had been recently freed from the Napoleonic terror:

On Saturday last the King was proclaimed throughout the City and Sunday was appropriated for the acclamation.—The Troops were reviewed by Marshal Beresford, yet never did I witness their going through their manoevres [sic] in such an inanimate manner:—never was such a Viva given by the Portuguese to their Sovereign; scarcely did one Soul open his mouth, excepting the Marshal and his Staff Officers:—it was a complete ‘Buonapartean Viva’ a forced shout of applause dying away in a groan. (11 April 1817)

Since most of the time he was consumed by work, it becomes difficult for the contemporary reader to detect such feelings of estrangement in the midst of commercial jargon and ledger accounts. He sought to be meticulous in his book-keeping and reports and sensitive to changes in market conditions, especially as far as fashion, trends, tastes and purchasing power went. He struggled to prove himself worthy of the trust and respect not just of his brother-in-law, but also of other foreign merchants who had already established their names in the Portuguese market. He even got carried away by the idea of opening his own establishment in order to fend off competition and to tackle the problem of low bids, which often forced him to keep the bales in store for unusually long periods of time.

In order to perceive how displaced he felt, one has to read between the lines. When his enthusiasm waned or his health gave way, an undeclared anxiety and irritation would surface. His less than flattering comments on Portuguese customs officials and the tone of his replies to his brother-in-law whenever suspicion of laxness or mismanagement hung in the air prove the point. He became impatient when ships from Brazil, New York or Falmouth were unduly delayed. He was unnerved by the negligence of long-standing debtors, who often turned a deaf ear to his entreaties. Besides, in spite of the considerable sums of money that passed through his hands, James was far from leading an easy and comfortable life. In a sense, it was through his own body that he first measured the degree of his maladjustment. He was constantly ill, poorly dressed, and found his lodgings uncomfortable. The weather did not suit him and he feared death might creep up on him. For some time he had to resign himself to ‘a Bed Room fitted up for me in the Warehouse, without any other convenience or sitting room’ (11 April 1817). He would wear the same clothes for months on end, winter and summer alike. Disease would take hold of him and he would be confined to bed for several weeks. His neat copperplate handwriting would then degenerate to illegible scribbling. In the spring of 1817 he would confess that ‘I have suffered very materially in my health since I came here’. Convinced that he was no longer fit for the job, he would then ask Thomas to let Ambrose Pollett, a friend of the family, replace him in the firm. His physical condition would not let him endure another winter in Lisbon. In his last letter, dated 29 November, he once more complained about his health, saying that the cold weather caused him to ‘spit blood in considerable quantities from the lungs’ and that he was afraid he would never be able to return to his homeland again ‘since I fell [sic] persuaded I shall never get better of the severe illness I had in the Spring of the year 1816’. To him Lisbon, thus, ended up representing the proximity of death, that ultimate moment of displacement. His fears, however, were unfounded and he went back to England where he remained in convalescence, before returning to Portugal. But once more the climate did not agree with him. His health worsened, especially after hearing the news of his nephew’s death in December 1818, and he was compelled to leave Lisbon one last time.[3]

In the course of his stay, James was badly in need of a focal point to keep things in perspective and letter writing served such a purpose. More than anything else, it allowed him to keep his sense of belonging alive. These letters ended up being the only bridge not just to his origins, but above all to his own identity. He felt so helpless when his sister failed to reply to his letters that ‘it even grieves me to the heart when I reflect upon it’ (17 February 1816). This sentimentality towards his family is in marked contrast with his attitude as an observer. Although Hutchinson cannot entirely detach himself emotionally from what he witnesses, there is a kind of Verfremdungseffekt in his writing, a journalistic objectification of the topics he covers, whereby the distance between himself and the other is never to be entirely spanned.

4. Prying into a Gentleman’s Private Correspondence: Issues of Intimacy, Confidentiality and Objectivity in Translation

Translating something as intimate and confidential as private letters has the potential to border on voyeurism. It raises issues that concern the ethics of translation, since the translator, unlike the casual reader, is supposed to leave no stone unturned in his struggle to reach communicative effectiveness. His labour consists in unveiling all meanings, in ransacking the secrets of the author’s mind, and, if necessary, in exposing the frailties of his body. The innermost thoughts are not fenced off from the translator’s dissecting tools. In this sense, translation is to be viewed as an act of intrusion and, simultaneously, of extrusion (in other words a disclosure and a close examination of that which pertains to the private sphere). The former constitutes a form of violation, of disrupting that which belongs to the realm of the confessional and becoming, to borrow the words of St. Augustine, ‘privy to the secrets of conscience’; whereas the latter manifests itself in the form of violence, destroying the integrity of the textual body, vivisecting it and exhibiting it to the public gaze. Nevertheless, such violence is mitigated by the transmutational properties of time. Over time, these texts have acquired the status of archaeological evidence, which does not necessarily mean that in this respect the position of the translator is less delicate. After all, he was not the addressee of the letters and that fact alone poses some problems. An outsider may find it difficult to penetrate the referential fabric of the letters. Unlike travel accounts or autobiographies written for publication, these texts were not intended for a wide readership. They were personal in tone and content, and the writer knew what responses to expect from his only reader living across the English Channel. The writer did not project an ideal or fictional reader to whom he might grant full right of access to the world recreated in his prose. As a consequence, his world remains sealed off from a larger audience and the translator is forced to break into the textual space like a trespasser. Implicatures lie hidden within this corpus of letters but they can never be entirely unravelled: whatever inferences the translator may draw, he or she will always lack the necessary background knowledge to establish their validity. Such implicatures, one must not forget, are a symptom of the close relationship existing between the two correspondents. Implicit meanings result from a common experience, excluding other readers. Fortunately, the text in question is generally far more objective and factual than one would suppose, and this alone gives the translator significant leverage over the hidden aspects of the correspondence. It is in the terrain of factuality and narrativity that the translator moves free from major constraints, although it is certain that the faithfulness of the representation can never be taken for granted (see Polezzi 2004: 124).

Of course one cannot expect to find in such letters a precise and exhaustive portrait of Beresford’s Lisbon, systematically organised in such a way as to cover all possible angles. What we get instead is a myriad of disparate images that can hardly be coalesced into one single picture. The reason is obvious: the stories he tells do not follow any thematic pattern, other than the fact that all of them revolve around the city itself. Apart from the town of Sintra, a popular tourist resort in the nineteenth century, where he spent some time ‘for the benefit of my Health which, thank God I have recovered beyond my expectation’ (14 June 1816), he never set foot outside of the capital (or at least there is no archival evidence of him doing so) and therefore he apparently did not know what was going on in the rest of the country. His letters lack the ‘horror and pity’ William Warre experienced as he crossed the country chasing after the fleeing French army and encountering ‘many people and children absolutely starving and living upon nettles and herbs they gathered in the fields’ (Warre and Warre 1909: 222). Not even Sintra, that ‘glorious Eden’ with its ‘views more dazzling unto mortal ken than those whereof such things the Bard relates’, as Byron wrote in his celebrated Childe Harolds Pilgrimage (1812), succeeded in enrapturing our author, who preferred to remain faithful to whatever notable occurrences Lisbon had to offer the outsider’s gaze.

Hutchinson’s short narratives appear scattered throughout the letters in a rather random way, and it is their reading as anecdotal collages, rather than as a set of tightly-woven, interrelated stories, that allows the reader to gain a taste of the spontaneity of the narration and the ingenuousness of the narrator. Although the anecdotal episodes themselves are self-contained and refer only to fragments of both individual and collective experiences in early nineteenth-century Lisbon, they play an important part in the process of historiographical reconstruction of the past. The historiographical value of the letters lies in the fact that they contain accounts that were neither censored nor doctored: no one ever scrutinised or edited the stories, which were simply committed to paper without any concern for accuracy, trustworthiness or factuality. The ensemble of letters forms a sort of scrapbook containing clippings or mementos that were never meant to be published. Such moments, however, were bound together by a common genetic code: they all emerged out of the drive for novelty, a drive partly explained by the way the processes of cultural displacement affected the author.

However, when it comes to Hutchinson’s values and ideological assumptions, they are not readily easy to detect. He preferred to position himself as an observer rather than as a commentator, and avoided getting entangled in elaborate considerations. If the translator wants to gain a glimpse of his ideas and opinions, then he/she must proceed by engaging in a symptomatic reading of the letters, observing, for example, the way he framed and skewed the subject matter, or how he got himself more or less emotionally involved with the events he narrated, or simply how he refrained from passing judgement on what he saw. Far from highly opinionated, the letters nonetheless give us the chance of peering into his personality, albeit obliquely.

Sometimes, however, he felt compelled to take sides, such as when he dared to air his own opinion on Beresford:

...being the weaker power & finding himself defeated in all his projects, it is reported that he is about leaving [sic] the Country, which in my opinion is the wisest step he can take, else a worse fate may attend him. (11 April 1817)

Such explicitness was rare. Shortly after the rebellion in Pernambuco, Brazil, Hutchinson censured himself for letting slip his views on the political turmoil that had gripped the country and decided to not to return to the issue for fear of reprisals:

You are well aware that it is necessary to be very cautious how we treat upon political subjects in this Country, for which reason I avoid any thing of this nature, only sofar [sic] as I suppose it may be connected with the interests of Mercantile Affairs. (4 July 1817)

His fears over the consequences of political dissent were not wholly misplaced. The horrific hanging of the Conspirators he watched on 22 October 1817, shortly before his departure, left a lasting impression on him:

[C]uriosity led me to be one of the spectators of this awful scene & however disgraceful hanging may be in England I can assure you it is not less so here. The Executioner is obliged to ride astride the shoulders of every man he hangs.—It was about four O’Clock in the Afternoon when the Prisoners arrived at the foot of the Gallows & was about midnight when this melancholy scene closed.—After the Execution of all 7 out of the 11 were burnt on a Funeral Pile on the spot.

Here, his voyeurism matched his horror as he came to the full presence of death—that dark character that kept resurfacing in his writing.

5. Translation as a Double Disjuncture

As we have seen, what was once private acquires, over time, an archaeological value: the status of artefact is conferred on language as privacy metamorphoses into historical evidence. In translation, chronological distance is of the essence: one might even argue that every translation has embedded in its genes an indelible anachronism. In sharp contrast with our contemporary world, where synchronous forms of communication and instantaneous access to information seem to have taken hold of the way we communicate with each other, the art and craft of translation necessitates the slow transit of time. It is a painstaking process of problem-solving, reflection and maturation. It takes time and perseverance. And when it involves the representation of past historical phenomena, as in the present case, the temporal dimension acquires critical significance. On the one hand, the translator cannot help excogitating his own condition as a historical subject: he becomes conscious of the relativity of values, of the differentials separating lifestyles, habitus (in the Bourdieusian sense) and Weltanschauungen. On the other, the target text ends up constituting the representation of a representation and, as such, it is, as Althusser once stated of ideology, a representation of an ‘imaginary relationship of individuals to their real conditions of existence’ (Althusser 1971: 162). And here, in the translation process, the time gap separating source and target texts functions not so much as a thread linking both acts of writing along a historical continuum but rather as a lens, generating several simultaneous optical effects, where light shifts in unsuspected ways and where appearance must be understood in its composite and elusive nature. The world of the (author’s) ‘present’ can never be reconstructed as such in the target text. The translator necessarily operates in the time gap between two ‘presents’ (his/her own and the author’s). That is why the translator’s labour must be that of a conscious re-representation of history. This, of course, entails much scrupulous work of detailed historical research, as well as the ability to articulate it within the translational process.

The crux of the matter lies in being able to dwell in the interstices between two languages, two cultures and two historical periods. This is the translator’s privilege and the source of many of his tribulations. To be able to lay claim to the ability to contemplate the insurmountable differences that separate not only languages but also cultures, one is required to perceive how far one’s own consciousness depends not only on λόγος and on the chains of meanings that help one make sense of the world, but also on the points of rupture of discourse, those points where signifiers and signifieds (regardless of the language) can no longer encompass those phenomena that keep resisting appropriation, including the culture of the other. In other words, one must learn to come to terms with the undecidability which undermines the certainties offered by our ingrained logocentrism.

As the translator shifts, in the course of the translation process, from one logosphere (in the Barthesian sense) to another, he realises that the movement itself does not (actually, cannot) entail the loss or gain, subtraction or addition of meanings. Meaning does not constitute some sort of universal currency (that is, manifestations of a universal language common to all human beings) that can be subjected to a process of direct exchange or transaction. Meanings cannot migrate freely from one language to another. I can only subtract meanings within the system they belong to. Languages weave their own networks of meanings and the exact value of each meaning, if it can ever be assessed, is to be determined only symptomatically by the effects generated by its presence or absence in one particular social and cultural context. To believe in the transferability of the meaning and its capacity to survive as a whole in two distinct linguistic and cultural environments (as in a process of ecesis) is not to realise something that Derrida pointed out: that even within the same language meanings not only differ (a problem of spacing), but are forever deferred (which is the condition of their temporality). One of the main problems of translation, therefore, is not just spatiality but also temporality, particularly the historical condition of the texts.

And this, I think, poses an obstacle far more difficult to overcome, since it has to do with the impossibility for the translator to render two externalities compatible in one single (target) text. Just as Hutchinson was compelled, as an expatriate, to come to terms with the social and cultural reality of his host country[4] (which is, for all purposes, a question of spatiality), so the translator, like a migrant travelling through time, is forced to come to grips with an ancient world governed by laws long forsaken and now irretrievable (the question of temporality). And since both writer and translator are forever barred from a fully unmediated contact with the unconsciously lived culture of the Other, both seeing it as something external to themselves, though not necessarily negative, their attempts to assimilate cultural elements and national idiosyncrasies can only take place on the terrain of the imaginary, which enables them to crop, select, filter and reshape elements and idiosyncrasies in order to discursively tame the otherness. It is when the translator is trying to tackle texts of this nature that he feels – to allude to one of Derrida’s most quoted metaphors, borrowed from Shakespeare – that ‘time is out of joint’, namely that he is supposed to take up the writer’s voice, but without being able to adjust either to the discursive and ideological framework within which the texts once gained their coherence, or to the past ‘structure of feeling’ (to use one of Raymond Williams’s concepts of cultural analysis) that informed the emotions, thoughts and actions of the original writer (Williams 1965: 64-6).

Translators of travel writing therefore have to operate on a double disjuncture. On the one hand, they have to deal with the cultural gap that exists between the author and the people he visits (Hutchinson and the Portuguese), a gap which over-determines the perceptions, constructs, responses and projections of otherness of the British expat, but which -- since it is barely made explicit in the text -- can only be detected by means of a symptomatic reading. On the other hand, translators have to negotiate the disjunction that will always separate them from the time and the concrete conditions under which the texts saw the light of day -- a disjunction that is further amplified by the impossibility of mapping the exact location of the intersection of cultures which gives the letters their characteristic intercultural tension (see Cronin 2000: 6). Therefore, the translator is left with no choice but to try to overcome these two disjunctions, both of which constitute distinct moments of resistance to interpretation.

The translator’s path is strewn with obstacles, for the minute he or she starts translating the text that distinction is no longer clear: the two moments overlap and the barriers between them become blurred, since his or her gaze is constructed in and through the gaze of the expatriate. How can we then circumvent the limitations to translation that such a double disjuncture imposes? Of course a careful, detailed investigation into the empirical elements offered by the letters and the issues broached therein must always be conducted, but this is not enough: it can only be through a critical awareness of these tensions and resistances that translators may decentre themselves and avoid the pitfalls of identification and idealisation. It is this decentring at the core of translation that ends up being in itself a form of travelling. After all, ‘translatio’ in Latin means ‘carrying across’, ‘transporting’, ‘transferring’, and, in contrast to what we may think, it is not the source text that is ‘carried across’ to a target culture. It is rather the translator and his reader who are invited to venture across a frontier -- the frontier that sets the limits to their identities, values and representations, and that is both spatial and temporal.

In fact, the main challenges to the translation of these letters were posed by the problem of temporality, that is, by the difficulties of bridging the time gap. The first issue to be tackled was the stylistics of the Portuguese target text. It was not just a matter of finding the best equivalents and transferring contents from the source text into the target language without major semantic losses. It was also a matter of finding a style and a register that could somehow match the original ones. In order to do that, I compared the letters to similar archival and bibliographical sources in Portuguese. Two manuals of commercial correspondence proved invaluable: Arte da correspondência commercial ou modelos de cartas para toda a qualidade de operações mercantis [The Art of Commercial Letter Writing or Letter Templates for all Sorts of Trade Operations] (Anon.; 1824) and Monlon’s Arte da correspondência commercial ou escolha de cartas sobre o commercio [The Art of Commercial Letter Writing or a Selection of Business Letters] (1857), the only key style manuals of the day in this area still available for consultation in the Portuguese National Library. The analysis of the examples of letters allowed me to determine the way in which the target text was to be drafted.

One of the most complicated aspects I had to deal with was choosing the mode of address: the original letters invariably start with ‘Dear Brother’, and then the addressee is always referred to with the second person personal pronoun ‘you’. In Portuguese, this is not so linear. In the early nineteenth century, modes of address would have varied according not only to social class, age or degree of familiarity, but also to written language conventions. ‘You’ could be translated either as ‘Tu’ (too informal; the verb is conjugated in the second person singular), ‘Você’ (slightly more formal; the verb is conjugated in the third person singular), ‘Vossa Mercê’ (idem), or ‘Vós’ (more formal; verb conjugated in the second person plural), among several other possibilities. Back then, a relationship with a brother-in-law, close as it might have been, did not necessarily imply the use of the informal ‘tu’, since informality and closeness are not synonyms. The way Hutchinson closed the letters (‘Your ever Affectionate Brother’) bears witness to such emotional proximity, but it is far from being indicative of a relaxed, informal manner. The solution to the difficulty in ascertaining whether we were dealing with informality or politeness was partly given by the 1824 manual. The plural ‘Vós’ is used when addressing both singular and plural persons, but in some cases all we have is the initial ‘V—’, which could stand either for ‘Vós’, ‘Você’ or ‘Vossa Mercê’. When the ‘V—’; form occurs, the verb is conjugated in the third person singular, midway between formality and affable politeness. This was the form I resorted to throughout.

Another difficulty had to do with wording. The manuals proved useful in guiding my lexical choices. I wanted to give the translation a distinctive period flavour to represent the historical dimension of the original letters. For example, ‘company’ could be translated either as ‘sociedade’ or ‘empresa’, but these words barely appear in the 1824 manual, especially when referring to one’s own company. Instead, the commonest word is ‘caza’ [House] sometimes ‘caza de commercio’ (dated spelling), which I decided to adopt. Many more old-fashioned or outdated Portuguese words that appear in the manual were likewise retrieved: ‘embolço’ [imbursement]; ‘estimar’ [to believe; to guess];  ‘fazer-se de vella’ [to set sail]; ‘governo’ [management]; ‘sortimento’ [assortment]; ‘sortir’ [to sort; to provide]; ‘praça’ [exchange or financial centre; market]; ‘rogar’ [to beseech]. The manual was equally useful in providing formulaic language that was pretty close to some passages in Hutchinson’s letters: ‘Sacámos hoje sobre vós pelo importe da factura (…) L... a 60 dias á ordem de…’ [Today we drew on you for the sum of £… at sixty days]; ‘Vosso reverente servidor’ [Your very Obedient Servant]; ‘Por esta confirmamos a nossa circular de (…) desde a qual ainda não tivemos a satisfação de receber alguma vossa…’ [Without any of your Favors since mine of the … I have now to inform you…].

Another challenge was related to the commercial jargon both in English and in Portuguese. Nowadays commercial terminology in both languages is much more complex, but most of the neologisms that currently exist in Portuguese are English words. Back then, that influence was more tenuous. In any case, the search for the right equivalent would have always been time-consuming. ‘Bill’ alone, for instance, could be equivalent to as many things as ‘letra’, ‘letra de câmbio’, ‘saque’, ‘promissória’, ‘papel comercial’, ‘título de comércio’, ‘factura’, or ‘facturação’. If we multiply this by the wide spectrum of nomenclatures related to those areas of economic activity Hutchinson was directly or indirectly involved in, we have an idea of the complexity of the task.

To start with, there were the inner workings of the wool trade business. I had to unwind the ball of yarn of the English wool and worsted industry, including all the details concerning the different stages of the manufacturing process: recognising the provenance and differences in quality of the raw wool available in both the Portuguese and Spanish markets, the various patterns of the warp and weft, the way the cloth should be cut or dressed, specific types of woollen cloths, their designs and colours, and so on. One particular stumbling block was the enigmatic ‘37 R., 6 F., 4 S., 1 T. & 11 A.’ (letter dated 9 August 1816). It took me a while before I learnt from a magazine published in London in 1804 (Tilloch 1807: 239-42) that the initials did not stand for any English or Portuguese words, but for Spanish ones. They referred to the way Spanish wool (which also included Portuguese wool) was classified: Primera or Refina (R.), Fina (F.), Segunda (S.), Tercera (T.) and Añinos (A.).

Moreover, since conducting business ventures overseas back then was not without its risks, I had to acquaint myself with the idiom used in cargo and shipping insurance, learn about risk-assessment, shipping deadlines, storage conditions, bills of lading, types of merchant ships crossing the Atlantic, and so on. But then there are also taxes and duties, customs procedures and the requirements of port authorities, the valuation of the bales in the Cocket,[5] goods lodged at the Custom House not yet dispatched -- all of this wrapped up in a language of its own, which has to be patiently disassembled, explored, digested, and then reassembled and fine-tuned in the translation process. In order to penetrate that language I had to resort to historical research once more. I visited the ‘Torre do Tombo’ (the Portuguese National Archives) and consulted the records from the customs houses that existed in Lisbon at that time: the ‘Alfândega Grande do Açúcar’, the ‘Alfândega das Sete Casas’, the ‘Alfândega da Casa dos Cinco’ and the ‘Casa da Índia’, the first of which provided invaluable information about the duties on wools and worsted, the classification of wools and of all sorts of cloths, their quantities and provenance, and so on. In the records of the ‘Casa da Índia’, the inventory of the cargo of the French ship Le Commerciant [sic], seized in the summer of 1809, reminds us of the risks faced by merchants like Hutchinson.

I adopted a domesticating approach to a certain extent, adding explanatory footnotes whenever words, phrases or referents might challenge the modern reader’s understanding of the target text. However, since the Revista de Estudos Anglo-Portugueses is aimed at a scholarly readership, it proved unnecessary to insist on the explanation of cultural or linguistic aspects that they are supposed to be already acquainted with. Differences in style between early nineteenth-century and early twenty-first-century Portuguese are noticeable, but they do not make the text less intelligible. In any case, stylistic conventions should not pose a problem for all the scholars who are used to working with documents of that period. So I kept the footnotes to a minimum. The future publication of a book containing the complete correspondence of the Farrer family, this time aiming at a more general readership, will entail a different explanatory methodology, but not a different stylistic treatment.

6. Conclusions

Writing narratives of displacement and travel is in itself a translational act, where the author is always seeking to translate into his mother tongue the manifestations of the culture of the other.[6] The translator of travel writing, in turn, operates on a double disjuncture – the gap between the author and the visited culture, on the one hand, and the gap between the translator and the author, on the other – threefold if you include the inevitable temporal disjuncture. In the process, the translator is forced to question his identity, values and the representations of his own nation and people, especially if the original text is non-fictional and therefore stakes a claim to the immediacy and truthfulness of the experience. The translator thus has to achieve a tour-de-force in bridging all three gaps and rendering the text accessible to the contemporary reader. However, the meanings in the target text will always have but a spectral relation with the ones in the source text: they are constructed at the same time as a re-apparition of a former presence (that does not present itself as full presence) and as the apparition of a new presence –a new text in its own right. This distance between the source and target texts becomes more difficult to span when historical time – fissured as it has been, in this particular case, over these past two centuries by sudden ruptures and discontinuities – keeps eroding the paths that could render the source text recognisable to the reader: hence the importance of the translator’s historical consciousness and the necessity of articulating historical research with the translation process, since any translation of historical material that disregards the intelligibility of historical processes lacks the authority to stake claims to accuracy and credibility.


Althusser, Louis (1971) Lenin and Philosophy and Other Essays, trans B. Brewster, London, New Left Books.

Bethell, Leslie (1984) Colonial Brazil, Cambridge, Cambridge University Press.

Borges de Macedo, Jorge (1963) Problemas da História da Indústria Portuguesa no Século XVIII, PhD diss, University of Lisbon, Portugal.

Casas Pardo, José (ed.) (1992) Economic effects of the European expansion, 1492-1824, Stuttgart, Steiner Verlag.

Cronin, Michael (2000) Across the Lines: Travel, Language, Translation, Cork, Cork University Press.

Fletcher, J. S. (1919) The Story of the English Town of Leeds, New York, Macmillan.

Gentzler, Edwin (1993) Contemporary Translation Theories, Clarendon, Multilingual Matters.

Glover, Michael (1976) “Beresford and His Fighting Cocks”, History Today 26, no. 4: 262-8.

Lopes, António (2009) “Cartas inéditas de um jovem burguês 1815-1817” (1.ª parte) [“Unpublished letters of a young bourgeois 1815-1817” (1st part)], Revista de Estudos Anglo Portugueses, no. 18: 93-133.

--- (2010) “Cartas inéditas de um jovem burguês 1815-1817” (2.ª parte) [‘Unpublished letters of a young bourgeois 1815-1817’ (2nd part)], Revista de Estudos Anglo Portugueses no. 19: 175-204.

Maxwell, Kenneth (2004) Conflicts and Conspiracies: Brazil and Portugal, 1750-1808, London, Routledge.

Newitt, Malyn (2004) Lord Beresford and British Intervention in Portugal, 1807-1820, Lisbon, Imprensa de Ciências Sociais.

Pijning, Ernst (1997) “Passive resistance: Portuguese diplomacy of contraband trade during King John V’s reign (1706-1750)”, Arquipélago – História 2, no. 2, 171-191.

Polezzi, Loredana (2004) “Between Gender and Genre: The Travels of Estella Canziani” in Perspectives on Travel Writing, Glenn Hooper and Tim Youngs (eds), Aldershot, Ashgate: 121-37.

Tilloch, Alexander (1807) The Philosophical Magazine: Comprehending the Various Branches of Science, the Liberal and Fine Arts, Agriculture, Manufactures and Commerce. vol. 27. London, R. Taylor.

books.google.pt/books?id=fp9JAAAAYAAJ&printsec=frontcover&source=gbs_ge_summary_r&cad=0#v=onepage&q&f=false (accessed 15 April 2011)

Warre William, and Edmond Warre (1909) Letters from the Peninsula, 1808-1812, London, John Murray.

Williams, Raymond (1965 [1961]) The Long Revolution, Harmondsworth, Penguin.


[1] Ref. No. E 140/34/1. Records of the Exchequer: King's Remembrancer: Exhibits: Farrer (and another) v Hutchinson (and others). Scope and content: Letters to Thomas Farrer from his brother-in-law, James Hutchinson (Jnr.), in Lisbon. Covering dates: 1815-1817.

[2] Manuel J. G. de Abreu Vidal. Análise da sentença proferida no juízo da inconfidencia em 15 de Outubro de 1817 contra o Tenente General Gomes Freire de Andrade, o Coronel Manoel Monteiro de Carvalho e outros... pelo crime de alta traição. Lisboa, Morandiana, 1820; José Dionísio da Serra. Epicedio feito, e recitado em 1822 no anniversario da sempre lamentável morte do General Gomes Freire de Andrade. Paris, 1832; Joaquim Ferreira de Freitas. Memoria sobre a conspiraçaõ [sic] de 1817: vulgarmente chamada a conspiração de Gomes Freire. London, Richard and Arthur Taylor, 1822.

[3] He outlived Thomas (who died circa 1820) and was appointed executor of his brother-in-law’s estate.

[4] A process E. Gentzler (1993: 37) calls ‘domestication’.

[5] A customs office in Britain where detailed records of exports were kept.

[6] On the relation between travel and translation see Lesa Scholl (2009) “Translating Culture: Harriet Martineau’s Eastern Travels” in Travel Writing, Form, and Empire: The Poetics and Politics of Mobility, Julia Kuehn and Paul Smethurst (eds), London, Routledge; Susan Bassnett and André Lefevere (1998) Constructing Cultures: Essays on Literary Translation, Clevedon, Multilingual Matters; and Susan Bassnett (2002) Translation Studies, London, Methuen.


About the author(s)

Antonio Manuel Bernardo Lopes, PhD in English Culture, MA in Anglo-Portuguese Studies (specialty in English Literature) and BA in Modern Languages and Literatures
(English and German), is Senior Lecturer (Professor-Adjunto) in English Studies with the Department of Languages, Literatures and Cultures at the School of Education
and Communication, University of Algarve, where he teaches English language, literature and culture, literary analysis and supervises ELT postgraduate projects. He is
also the director of studies of postgraduate programmes in ELT and translation. He is a researcher at the Centre for English, Translation and Anglo-Portuguese Studies
(FCHS/UNL and FLUP), working with the following research groups: Anglo-Portuguese Studies; Literature, Media and Discourse Analysis; British Culture and History. He
has also participated in several European-funded projects related to teacher training and computer-assisted language learning. He is currently the EUROCALL
representative in Portugal. His doctoral dissertation is entitled The Last Fight Let Us Face: Communist Discourse in Great Britain and the Spanish Civil War.

Email: [please login or register to view author's email address]

©inTRAlinea & António Lopes (2013).
"Translating Echoes Challenges in the Translation of the Correspondence of a British Expatriate in Beresford’s Lisbon 1815-17"
inTRAlinea Special Issue: Translating 18th and 19th Century European Travel Writing
Edited by: Susan Pickford & Alison E. Martin
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/1967

Cfp: Translation meets Book History: Intersections 1700-1950

By The Editors


©inTRAlinea & The Editors (2017).
"Cfp: Translation meets Book History: Intersections 1700-1950"
inTRAlinea News
Edited by: {specials_editors_news}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2267

©inTRAlinea & The Editors (2017).
"Cfp: Translation meets Book History: Intersections 1700-1950"
inTRAlinea News
Edited by: {specials_editors_news}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2267

The Martha Cheung Award for Best English Article in Translation Studies

By The Editors


©inTRAlinea & The Editors (2017).
"The Martha Cheung Award for Best English Article in Translation Studies"
inTRAlinea News
Edited by: {specials_editors_news}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2266

©inTRAlinea & The Editors (2017).
"The Martha Cheung Award for Best English Article in Translation Studies"
inTRAlinea News
Edited by: {specials_editors_news}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2266

Quality aspects in institutional translation

By Lucja Biel (University of Warsaw)


©inTRAlinea & Lucja Biel (2017).
"Quality aspects in institutional translation"
inTRAlinea News
Edited by: {specials_editors_news}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2265

About the author(s)

Łucja Biel is a linguist and translation scholar (Assistant Professor at the Institute of Applied Linguistics, University of Warsaw, Visiting Lecturer at City University
London), sworn translator, deputy editor of the Journal of Specialised Translation and Secretary General of the European Society of Translation Studies.

Email: [please login or register to view author's email address]

©inTRAlinea & Lucja Biel (2017).
"Quality aspects in institutional translation"
inTRAlinea News
Edited by: {specials_editors_news}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2265

Englishing A Spanish Romance

Translating Spanish Rivalry into English Patriotism in Margaret Tyler’s Mirror of Princely Deeds and Knighthood

By Rachel Roberts (North Greenville University, USA)

Abstract & Keywords

Margaret Tyler’s Mirror of Princely Deeds and Knighthood (1578) translates Book One of Espejo de Príncipes y Cavalleros by Diego Ortúñez de Calahorra (1555). The Mirror went through three English editions by 1599, creating an English appetite for the Continental romance. This essay argues that Tyler’s success is no accident but, instead, is a tribute to her choice of subject and skill in translation. Throughout her generally accurate translation, Tyler alters, clarifies, and intensifies Ortúñez’s language in order to produce a more orderly—and more English—tale. Thus, Tyler participates in the medieval and Renaissance ideal of translatio (translatio studii, translatio imperii), the transfer of cultural knowledge through linguistic translation. This essay examines how Tyler frames both the Mirror’s Spanish origins and its new English identity under her care. Throughout close readings of Tyler’s preface and text, this essay demonstrates that Tyler deliberately evokes both foreign and national identities in order to appeal to readers as well as to demonstrate her own skill as a translator.

Keywords: literary translation, history of translation, translation and gender, early modern literature, Margaret Tyler

©inTRAlinea & Rachel Roberts (2017).
"Englishing A Spanish Romance Translating Spanish Rivalry into English Patriotism in Margaret Tyler’s Mirror of Princely Deeds and Knighthood"
inTRAlinea Volumes
Edited by: {specials_editors_volumes}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2264


Margaret Tyler’s Mirror of Princely Deeds and Knighthood (1578) translates Book One of Espejo de Príncipes y Cavalleros by Diego Ortúñez de Calahorra (1555).[1] This chivalric romance follows the exploits of the Greek Emperor Trebatio and his sons, The Knight of the Sun and Rosicleer. Tyler’s translation proved successful both in its own right and in its influence upon other translators of romance. The Mirror went through three editions by 1599; contemporary references to the Mirror in the works of John Lyly and Ben Jonson suggest widespread familiarity with Tyler’s text (Boro 2014: 4).[2] In a decade when Anglo-Spanish tensions ran high, Tyler’s work catapulted the Spanish romance to dazzling literary success.

The success of Tyler’s Mirror in England is no accident but, instead, results directly from Tyler’s choice of genre, choice of subject, and approach to translation.  The romance was, as Helen Cooper demonstrates, “the major genre of secular fiction” in England during the medieval and early modern periods (2004: 2); Tyler’s translation of the Mirror was the first in a wave of Iberian romances to sweep across England, including the popular Amadís de Gaula (Boro 2014: 4). Tyler’s prefatory epistle “To the Reader” emphasizes that her Mirror fulfills the twin demands for “profit and delight,” demonstrating that she is aware of larger conversations about literature as well as romance (1578/2014: 49.4). Tyler’s preface and her dedicatory letter to Thomas Howard, Earl of Suffolk (1561-1626) further demonstrate her awareness of print culture and her own culturally fraught position as a woman writer. Despite evidence to the contrary, the Mirror proves ripe for English interest, containing English characters and settings that figure prominently in Tyler’s translation. In fact, Tyler alters, clarifies, and intensifies Ortúñez’s text in order to produce a more English tale, adding both vocabulary and cultural references that would be familiar and appealing to English readers (see Boro 2014: 3). Tyler thus participates in the medieval and Renaissance ideal of translatio, the transfer of cultural knowledge through linguistic translation.[3] In fact, in her preface “To the Reader,” Tyler describes her work as “Englishing this author” (1578/2014: 49), demonstrating that she is aware of the tensions inherent in her choice of a Spanish work while she creates a uniquely English experience for her readers.

In this essay, I examine how Tyler frames both the Mirror’s Spanish origins and its new English identity. Throughout my close readings of Tyler’s preface and text, I argue that Tyler deliberately evokes both foreign and national identities in order to appeal to readers, to demonstrate her skill as a translator, and to frame her own identity as a woman translator. In the first section of this essay, I examine Tyler’s preface, demonstrating how Tyler uses her paratextual material to negotiate the fraught relationship between Spain and England as well as to establish her own position as a woman in early modern print culture. Next, I explore Tyler’s textual adaptations: in the paper’s second section, I cover Tyler’s treatment of the characters Briana and Edward, while in the third section, I closely read scenes involving the character Olivia as well as an important tournament set in England. Tyler espouses translatio throughout the text, altering the Espejo’s vocabulary, tone, and characterization in service of her English readers. Tyler is particularly attentive to the concerns of female characters (and, perhaps, female readers by extension) as well as to English national pride—“Englishing,” as she calls it, the Mirror in order to comment upon these aspects of her own culture. By closely examining Tyler’s methods throughout the work, I posit that the Mirror’s success with English readers was a planned consequence of Tyler’s choice of subject as well as her Anglicized translation methods.

Although Tyler’s Mirror was designed to appeal to her sixteenth-century English audience, modern readers and scholars have ignored the work until recently. In fact, in the first modern edition of Tyler’s romance, Joyce Boro only lists seven works under “Further Reading” (2014: 37). This lack of attention is all the more surprising because of the Mirror’s place in literary history: Tyler is the only sixteenth-century Englishwoman to publish a romance, either an original or a translated one, making the Mirror an important point in the history of women’s writing in English (Coad 1996: ix). In addition, Tyler is the first author to translate a romance directly from Spanish, a decision that profoundly affected English romance (Boro 2014: 3, 6). Despite these accomplishments, scholars are often uncomfortable assigning a creative status to translators; even Uman and Bistué, who reframe translation as “collaborative authorship,” claim that “Tyler’s translation does not offer us a glimpse of the author’s imaginative flair or poetic skill” (2007: 298). Such marginalization of translation means that authors like Tyler, whose only extant work is a translation, are often ignored except in studies specifically focusing on translation or, for Tyler’s Mirror, in studies of the romance genre.[4] The few exceptions to this rule include works about Tyler’s preface, which is often studied or anthologized separately, and the groundbreaking work on Tyler by Tina Krontiris and Louise Schleiner.[5] Now that Boro’s edition makes Tyler’s Mirror more accessible, more readers and scholars should make this text an object of study, both for its contribution to literary history and—as in the present essay—for Tyler’s linguistic and artistic craft.

The Mirror as Spanish: Tyler’s Dedication and Preface

In sixteenth-century England, a Spanish work seems an unlikely choice for popular success. In 1578, the year of the Mirror’s publication, England and Spain were embroiled in political and religious conflict (building up to the infamous Spanish Armada of 1588).[6] Rather than shy away from the Mirror’s dangerous origins, Tyler deliberately acknowledges the Mirror’s Spanish origins throughout her prefatory material. In her address “To the Reader,” she writes, “The first tongue wherein it was penned was the Spanish, in which nation, by common report, the inheritance of all warlike commendation hath to this day rested” (Tyler 1578/2014: 49). This reference to the “warlike” nature of Spain could be either compliment or criticism—suggesting that Spain is superior to England in military might or (through the use of “to this day”) offering hope that England will now be ascendant. Tyler’s initial reference to the Mirror’s Spanish origins thus suggests a context of political and military rivalry that seems unsuited for literary enjoyment.

Inasmuch as a Spanish romance is an odd choice for sixteenth-century England, Tyler herself is an unlikely figure for success, given the fraught cultural conversations around women as writers and readers of romance. The romance was seen as particularly “problematic” for women, since its focus on love and war was deemed inappropriate (Arcara 2014; see also Krontiris 1988: 23-24). Despite such theoretical problems with romance, such works were frequently circulated, printed, and read in early modern culture, by men as well as women (see Hackett 2000: 5-9, 70-75).[7] Tyler touches on the fraught romance genre in her preface “To the Reader,” emphasizing the Mirror’s moral uprightness and suggesting that the work “may bring thee [readers] to a liking of the virtues here commended” (1578/2014: 49).[8] Tyler’s later mention of the “delights” and “profitable reading” (1578/2014: 51) available through the Mirror also evokes, as Arcara (2014) notes, the Horatian and Aristotelian traditions later endorsed in Sir Philip Sidney’s “Defence of Poesy.” Tyler’s epistle “To the Reader,” then, emphasizes the positive, exemplary values of the romance to convince her readers that, despite the work’s genre, it is worth reading.

While Tyler might be suspect as a woman writing romance, she is perhaps less controversial as a translator. Translation itself was not problematic for women writers, and translations were often dedicated to women; in fact, the Renaissance translator John Florio famously remarked that “all translations are reputed femalls.” As Arcara (2014) describes, liberal humanists such as Juan Luis Vives viewed translation of properly moral works as an appropriate intellectual exercise for educated women (see also Agorni 1998: 181). Many early modern women, including such well-known figures as Queen Elizabeth I, Mary Sidney Herbert, Anne Vaughan Locke, and Margaret More Roper, translated a variety of works.[9] The propriety of translation did not, however, extend to publication; Krontiris notes that “to publish meant to engage in public self-display” (1988: 21). Uman and Bistué further note that publication brings the woman writer into “the public sphere of literature” where her “chastity is rendered suspect” because her work is promiscuously available to all readers (2007: 302). Of course, such suspicion did not prevent women translators from publishing their work (see Clarke 2001: 3). Tyler boldly publishes her Mirror, but her lively defense in “To the Reader” demonstrates her awareness that readers will not only find the work’s foreign origins suspect—they may be wary of romance in general and a romance published by a woman in particular.[10]

In other words, in order for Tyler’s Mirror to achieve success with its English audience, Tyler must carefully contextualize its foreign origins, her own role as translator, and her audience’s gendered and political expectations. She does so through a variety of methods in her two prefatory pieces, the dedication to Thomas Howard and the epistle “To the Reader.” One primary way in which she balances these issues is through the use of the modesty topos, deliberately downplaying her own agency (see Boro 2014: 27). Her choice of Thomas Howard as a dedicatee, for instance, allows her to frame modesty in gendered terms. Rather than dedicating her work to a woman—a practice she discusses in “To the Reader”—Tyler chooses a respectable male figure. Tyler emphasizes her deep devotion and faithful service to Howard’s family before directly stating her need for Howard’s help: “Under your honour’s protection I shall less fear the assault of the envious” (48.28-29). Uman and Bistué refer to this rhetorical strategy as “don[ning] the gown of the lady in distress” (2007: 304); Tyler’s acknowledgement that she needs “protection” from “the envious” both humbles herself and exalts Howard. In addition, Tyler frames her work as an “exercise of translation” (1578/2014: 48.3), shifting any blame for questionable content to Ortúñez, the author, rather than Tyler, the translator. As Arcara (2014) argues, Tyler “devalue[s] the activity of translation” to defend “her gendered position as a translator of a secular text.”[11] Tyler thus employs the modesty topos in order to frame audience responses to her patron, her choice of subject, and her choice of translation rather than anything more controversial.

In addition to using the modesty topos to keep the work’s questionable origins from overwhelming her readers, Tyler maintains a deliberate neutrality on the fraught question of religion, which in sixteenth-century England was closely linked to questions of foreign powers—and to specific women. For Tyler’s first readers, Spain was the country of origin for Catherine of Aragon (d. 1536) and, through her daughter Mary (d. 1558), an entry point for Catholicism into the English monarchy.[12] In The Spanish Armada, Robert Hutchinson (2013: xii) presents the military actions of the Armada as “the climax to a war of religion, the Catholic Church versus the fledgling Protestant state of England.” Protestant-Catholic tensions were at the heart of the problems between these two countries, so Tyler’s religious allusions within an “Englished” Spanish text could be sites of great import (see McDermott 2005: 92). In her preface, Tyler deliberately sidesteps the question of religion; she explains that she chooses not to “write of divinity,” since she has never found “any book in the tongue which would not breed offence to some” (1578/2014: 50). This claim is an important part of Tyler’s self-reference as a woman writer, since writing on religious matters was usually less objectionable for women than writing on secular topics (Beilin 1997: xx). However, Tyler’s refusal to take sides on questions of “divinity” reveals her awareness of the problematic nature of her Spanish text for an English (Protestant) audience.[13]

Tyler’s chosen text thus carries many inherent difficulties for an English audience, from national rivalry to genre to gender to religion. Despite these difficulties, the Espejo is in fact an astute choice for an English adaptation. I posit, in fact, that Tyler chose the Espejo deliberately for its ease of adaptation to English cultural and literary sensibilities. The Espejo is much more Continental than it is Spanish as far as its characters and setting are concerned. The main character, Trebatio, is a Greek knight who secretly marries a Hungarian princess, Briana. Their bi-national twin sons, Rosicleer and the Knight of the Sun, travel in the prescribed romance manner to enchanted islands inhabited by giants. In addition to these varied nationalities, Espejo also features English characters: Briana is originally betrothed to an English prince, Edward, while the English king, Oliverio, along with his daughter, Olivia, feature in an episode where Rosicleer competes in an English tournament. Dangerously Spanish or not, the Espejo lends itself well to Tyler’s cultural adaptation.

Having chosen a tale that is ripe for English adaptation, Tyler also emphasizes positive aspects of the Mirror’s foreignness in her preface “To the Reader.” Tyler (1578/2014: 40) describes “that delight which myself findeth in reading the Spanish” among her reasons for translating the tale. Tyler’s own delight in reading the tale suggests an immediacy of experience that her translation replicates. In addition, as both Coad (1996: 8) and Krontiris (1988: 19) note, Spanish was not part of the typical English education; most writers who translated Spanish texts did so through an intermediary language such as French. By working directly from the Spanish, Tyler creates a more direct connection between her English reader and the “Spanish delight” awaiting them within the Mirror (1578/2014: 50). With these claims of delight and emphasis on her translation, however, Tyler also assigns herself an important place within the reader’s experience of the Mirror, for it is through Tyler’s mediation that this pleasure becomes available to English readers.

Tyler also emphasizes her important role in shaping readers’ experience through the references within her preface to the Espejo’s author, Diego Ortúñez. She refers to him several times, calling him “a stranger” (Tyler 1578/2014: 49) and “this Spaniard” (Tyler 1578/2014: 51). More specifically, Tyler frames her relationship to Ortúñez in terms of hospitality. She calls her translation “giving entertainment to a stranger, before this time unacquainted with our country guise” (Tyler 1578/2014: 49).[14] The virtue of hospitality has gendered implications that reflect on Tyler as a woman translating a male author; Schleiner (1994: 253n21) notes that hospitality was a “proper female image” and thus one of Tyler’s methods for feminizing her writing.[15] However, this framing of translation as hospitality is also vital to Tyler’s treatment of the Mirror as a Spanish work. Rather than pursuing conflict with a foreigner, Tyler hosts and welcomes this stranger. By referring to “our country guise,” however, Tyler also reminds her readers that she has dressed up her guest, so to speak, in English garments suitable for her readers to encounter and understand. Tyler is essentially a master of ceremonies, introducing her readers to a new and important acquaintance. This metaphor of hospitably welcoming the Espejo into English, decking it out to become the Mirror, thus brings together all of Tyler’s claims in the prefatory material. This metaphor at once emphasizes the text’s foreignness, places Tyler herself in a properly feminine position, and emphasizes her vital role as a translator who brings this pleasurable piece of literature to English readers for their “profit and delight” (1578/2014: 49).

The Mirror As English: Language and Characters

Having selected the Espejo with care for its resonance with English audiences, and having explored the text’s potential points of controversy, Tyler adapts the Espejo in order to advance her gendered and nationalistic purposes. First, Tyler draws readers’ attention to the concerns of female characters, represented by Princess Briana. Tyler thus suggests the importance of female characters with rich inner lives, a move which may be calculated to appeal to romance’s women readers.[16] Second, Tyler effects what Boro (2014: 1) calls a “national transposition” by using distinctively English vocabulary and by adding complimentary language to the Mirror’s portrait of the English Prince Edward. In doing so, she participates in the sixteenth-century surge in British nationalism most familiar from writers such as John Foxe, William Shakespeare, and Edmund Spenser (Maly 2003; Escobedo 2004). Through her use of literary techniques, vocabulary, and characterization, Tyler “Englishes” the Mirror for her readers—both men and women.

Tyler’s focus on Briana, one of the Mirror’s most prominent female characters, becomes obvious from her title itself. Tyler’s full title is The Mirror of Princely Deeds and Knighthood, Wherein is showed the worthiness of the Knight of the Sun and his brother, Rosicleer, sons to the great Emperor Trebatio, with the strange love of the beautiful and excellent Princess Briana, and the valiant acts of other noble princes and knights. Ortúñez’s Espejo, on the other hand, proclaims instead “the great chivalric deeds and very strange loves of the beautiful and excellent princess Claradiana” (Ortúñez title page; see also Boro 28).  Claradiana is an Amazonian princess who hunts and later fights with male knights; Tyler emphasizes instead the Hungarian princess, Briana. Both Briana and Claradiana appear in Tyler’s Mirror, but the substitution Briana in the title signifies Tyler’s greater concern for the princess. At first glance, this emphasis indicates a shift from a more radical woman (Claradiana) to a more traditional one. Briana is an obedient daughter (1578/2014: 62) and, later, a patient Penelope waiting for her missing husband’s return (226).[17] These are roles and attitudes that even the stodgiest sixteenth-century reader would think appropriate for a (real or fictional) woman.

However, Tyler’s emphasis on Briana’s complex inner life moves this initially stereotypical character into greater complexity. The conventional Briana is, as Krontiris demonstrates, portrayed in a way that is “sympathetic” and makes readers “aware of oppression” (1988: 67).  For instance, Tyler records Briana’s “fear” when Trebatio presses her to consummate their marriage (1578/2014: 64). Tyler also emphasizes the complex emotions Briana experiences at the birth of her sons: “The mother, full of pain with the travail which she had sustained as well as she could, laying them to her breasts, kissed and embraced them with such love and pity that the tears trickled down from her fair eyes” (1578/2014: 75). In such instances as these, Briana’s actions may be conventional for a romance heroine, but her emotionally complex portrayal is less so. Tyler’s substitution of Briana for Claradiana in the title, although it aligns with Briana’s traditional role in the text, also makes room for a less traditional, more complex portrait of Briana’s inner life. These adjustments to Briana’s character align with Tyler’s defense of herself as a woman reader and translator in her preface, and perhaps also appeal to the female readership of romance in general.

In addition to gendered changes, Tyler adapts the Mirror’s language to make the tale more English, appealing to patriotic readers. The most obvious way Tyler “Englishes” the text is through English literary techniques. For example, she employs alliteration that does not appear in the Spanish text, such as when she foreshadows Prince Edward’s defeat by noting that he is “ignorant of the sour sauce and woeful wedding which was in providing” (Tyler 1578/2014: 58). Tyler also includes metaphors and imagery that suggest English culture. For example, she adds theatrical imagery to the description of one of the Mirror’s many giants, calling him “rather…a tyrant in a tragedy than a jester in a comedy” (Tyler: 1578/2014: 150-151). Likewise, Tyler regularly uses the language of “humours” (1578/2014: 161), which Ortúñez does not employ. For instance, Tyler adds a reference to “melancholy” in describing the lovesick Trebatio (1578/2014: 63).[18] This language reflects the early modern (English) understanding of bodily humors and their role in emotions. All of these techniques and metaphors make the experience of reading Tyler’s Mirror—especially reading it aloud, as was common—particularly English.

In addition to these incorporations of imagery, Tyler (1578/2014) uses English vocabulary. Colloquial phrases such as “how now” (157), “hurly-burly” (167), and “God save you” (227) give the Mirror a particularly English cadence.[19] Tyler (1578/2014) also substitutes English measurements like the “bowshot” (66), “yard” (80), and “finger” (93) for Spanish ones. Boro (2014: 66n65) notes, for example, that where Tyler uses “bowshot,” the Espejo uses “trecho” (‘stretch’). Such vocabulary, spread throughout the Mirror, is not overwhelming. However, Tyler’s consistent addition of English vocabulary and imagery throughout the Mirror demonstrates her commitment to “Englishing” this work.

One specific use of English vocabulary has historical resonance. Tyler deliberately chooses the word “rover” to describe Mambriniano, a sea-faring villain. The term “rover” had many negative associations for the English reader in the 1570s—resulting from England’s sea battles with Spain. By choosing the term “rover” to describe Mambriniano (Tyler 1578/2014: 86, 87, 88), Tyler evokes the national tensions inherent in debates about pirates and privateers (see Boro 2014: 86n165). These tensions grew steadily throughout the sixteenth century as English privateers turned from French to Spanish vessels in their pursuit of wealth (McDermott 2005: 20-21, 29). In 1577, the year before Tyler’s Mirror was published, Elizabeth I made significant changes to the judicial system in order to prosecute pirates more effectively (McDermott 2005: 124). With such a campaign against piracy in England, it is not surprising that the “rover” Mambriniano is a sinister figure in Tyler’s Mirror. By picturing Mambriniano’s grim death at Florion’s hands, however, Tyler also suggests that pirates can be defeated (1578/2014: 87). Evoking, through the term “rover,” the current English piracy crisis, including the trouble these pirates caused in Anglo-Spanish relations, Tyler situates her translation within her historical moment; she also evokes the political tensions between England and Spain—potentially problematic ground for her English readers.

At times, Tyler’s use of English vocabulary emphasizes her own role in “Englishing” the work. For instance, in chapter 52, as Florinaldes is borne off the field after a defeat by the Knight of the Sun, the narrator admits, “it is uncertain whether [Florinaldes was] more grieved with the sore of his bruise than with the shame of his fall, so to be foiled before his mistress. But if I may meddle in school points, I think he had rather burst an arm than so to have cracked his credit with both lady and friends” (Tyler 1578/2014: 229). In this second sentence, which Tyler adds to the Spanish text, she makes her commentary particularly English by using the term “school points” (1578/2014: 229).[20] The Oxford English Dictionary (2015) records it as a single word, “schoolpoints,” referring to “a matter taught or debated in the schools.”[21] By naming this event a “school point,” Tyler suggests that Florinaldes’s emotions are open to debate—but Tyler also takes an interpretive side, reminding readers of her important mediatory role as translator.[22] Tyler’s addition of a distinctly English phrase to this commentary reinforces her role as translator while adding English phrasing to her text.

In addition to her particularly English vocabulary, Tyler also adjusts the Mirror’s attitude toward the English Prince Edward. While mentions of Briana reveal Tyler’s concern with women’s perspectives, Edward’s literary rehabilitation is clearly motivated by nationalism. In the Spanish text, Edward is an unattractive character whose role is to be literally replaced by Trebatio (who kills the English prince in order to wed Edward’s betrothed, Briana). Although Tyler does not change this sequence of events, she alters Edward’s characterization in positive ways. For example, Tyler (1578/2014) describes Edward as “strong, and valiant” (54) when Ortúñez merely calls him proud (Boro 2014: 54n33). Similarly, she translates the Spanish “sobervio” (‘haughty’) as “stout,” which has more positive connotations (Tyler 1578/2014: 59, Boro 2014: 59n46, Ortúñez 1555/1617: 7). Although Edward is ultimately defeated by Trebatio, Tyler stresses the English prince’s valor throughout their battle.[23] For instance, Trebatio is “so great and so big made that he seemed to be a giant” (Tyler 1578/2014: 59). Boro (2014: 11) notes that Tyler often “omits…references to the heroes’ size” so as not to conflate them with the evil giants more common in romance. In this scene, however, Trebatio appears to be a giant whom Edward tries to defeat, as any romance hero would. Edward is thus a valiant figure even though he dies at Trebatio’s hand. Furthermore, Trebatio is “disquieted” by Edward’s death and weeps over “the loss of so great a prince slain out of his own country in the beauty of his age” (Tyler 1578/2014: 60). Trebatio’s sorrow makes the Emperor more sympathetic, but it also acknowledges Edward’s greatness within this tale. Tyler’s rehabilitation of Edward, like her English vocabulary and her treatment of Briana, demonstrates Tyler’s commitment to translatio—to making the Mirror her (and England’s) own.

English Tournaments and Royalty: Princesses and National Pride

As the text progresses, Tyler moves from more subtle uses of translatio (vocabulary, character descriptions) to more overt statements of support for England. Such statements appear most fully in Tyler’s treatment of the English princess Olivia as well as in the depiction of a great tournament held in England.

Olivia, Edward’s sister, is the second major English character in the Mirror; Tyler consistently emphasizes Olivia’s role as the female heir to the British throne. Focusing on heirs and succession was difficult for the Elizabethan writer, although such topics lay within the purview of the romance (as in Shakespeare’s The Winter’s Tale). Plots revolving around succession grew increasingly tense as Elizabeth’s reign progressed and it became apparent that, in Cooper’s (2004: 353) words, “The heir to Elizabeth is not lost and awaiting recovery, but not there at all.” Succession was therefore a fraught subject even in fiction. Tyler’s treatment of Olivia, however, pays homage to England’s reigning queen by strengthening Olivia’s position as a future English monarch.

Earlier in the text, Tyler references English succession to emphasize the difference between English and Continental views on royal inheritance. As Briana and Clandestria discuss her new pregnancy (believing the baby’s missing father to be Prince Edward of England), Clandestria advises Briana, “if God give you a man-child…this your child is lawful inheritor of Great Britain in the right of his father, the king now living having no issue male” (Tyler 1578/2014: 73). Boro notes that Tyler adds the phrase “in the right of his father” in order to clarify the progression of inheritance, since on the Continent (Clandestria and Briana’s Hungary, Ortúñez’s Spain, or both), a male heir would be the only possible choice (2014: 73n104). However, Clandestria, a Continental interpreter, is wrong: England already has an heir in the person of Edward’s sister, Olivia. Tyler’s language with respect to Olivia reflects the British princess’s secure position as heir to the throne. Specifically, Tyler chooses the term “inheritrix” to describe Olivia’s position; for example, “She was brought up as being inheretrix to the state with great care by the king her father” (Tyler 1578/2014: 77; see also 139). Ortúñez’s term is “heredera,” literally “heiress” (1555/1617: 87). However, the Oxford English Dictionary (2015) records “inheritrix” as the most “technical” and formal of the related English terms inheritrix, inheritrice, and inheritress.  By choosing the legal term, Tyler emphasizes the legitimacy of Olivia’s succession.

In addition to this formal language, Tyler alters a conversation between Olivia and her father, decreasing the power assigned to Olivia’s future husband. Oliverio advises his daughter to accept her suitor, Don Silverio, “both for mine own liking and the common profit of my subjects” (Tyler 1578/2014: 235). In the Espejo, the text then explains that Olivia’s husband will “inherit and defend the kingdom” (“y que pueda defender este reyno y estado,” Boro 2014: 235n679, Ortúñez 1555/1617: 207). Tyler omits this emphasis upon the need for a king to defend England, avoiding the implication that Olivia’s husband would rule. Tyler is not so careful with all female heirs; in an earlier episode, Rosicleer declares a woman, Liverba, “mistress” of the Valley of the Mountains (Tyler 1578/2014: 141). He immediately declares it necessary “to match her with the chiefest inheritor of land and sea amongst them” (Tyler 1578/2014: 131). However, when the inheritance in question is the English throne—in Tyler’s day firmly occupied by Queen Elizabeth—Tyler avoids this preference for male rule. Oliverio may indicate a preference for his daughter to marry, but he does not suggest that the kingdom is jeopardized by her single status. By omission rather than addition, Tyler makes room for a royal female heir, reflecting England’s current reality.

Tyler’s translation choices with regard to Olivia include both gendered and political concerns, reflecting early modern England’s reality with a reigning queen as the head of state. Elsewhere, however, Tyler’s positive language regarding England is purely nationalistic in character. The most obvious use of such language is in chapter 32, which depicts a tournament held by King Oliverio of England. At this tournament, Rosicleer, one of the Mirror’s heroes, defeats his first giant and falls in love with Olivia. This scene is undoubtedly central to the Mirror’s plot; however, on its own, Rosicleer’s adventures at the tournament do not seem significant enough to justify Tyler’s painstaking alterations to nearly every sentence—by far the most concentrated revisions of the Spanish text in the whole Mirror. I argue that Tyler’s changes are so intense because this scene reflects Tyler’s overall hospitable and nationalistic purpose in its depiction of foreigners finding welcome in a triumphant England.

At one point, Tyler describes those present at the tournament as “both strangers and Englishmen” (1578/2014: 142). “Stranger” is the same term she used for the Espejo’s author in her preface “To the Reader,” where Tyler figures the Mirror as an act of hospitality (1578/2014: 49). In this scene, many foreign knights, many “strangers,” are welcomed to England. The tournament scene’s construction thus echoes the Mirror’s position as a foreign text brought into England; the scene’s language is also the fullest embodiment of Tyler’s translatio as she alters many phrases to provide more accurate descriptions of English culture as well as higher praise for her native land. Thus, Tyler’s cultural translation is at its height in the tournament scene, which provides a microcosm of Tyler’s project and must remain at the center of any discussion of “Englishing.”

The Mirror’s first portrait of England, in chapter 30, demonstrates the country’s need for chivalric rehabilitation. After years of searching for the “missing” Prince Edward, England is “very naked of able knights to defend it, whereas before it was best known in all the world for knighthood and chivalry” (Tyler 1578/2014: 133). As the king recalls his knights from foreign lands, Tyler adds to the Spanish text a description Oliverio’s “solemn triumphs” at finding his country once again “sufficiently furnished” of knights (1578/2014: 133, Boro 2014: 133n351). With the return of his English knights, Oliverio’s kingdom is once again restored to dignity. Although this passage ends by describing the grand tournament, Tyler’s text emphasizes that England is “sufficiently” (1578/2014: 133) restored with only its native knights. No foreign warriors are required. Tyler thus emphasizes England’s strength and sufficiency with her additions to the Spanish text.

Tyler continues this positive portrayal of England when she expands definitions of England’s wealth. First, Tyler (1578/2014) increases the value of the tournament prize: “a massy crown of gold, all set with pearls and precious stones, valued by all men’s deeming at the price of a great city” (133). Such a valuable prize enhances England’s reputation (Boro 2014: 133n352). Similarly, King Oliverio is willing to pay “more than London is worth” (Tyler 1578/2014: 142) to rid himself of an interfering giant. Tyler adds the reference to London and its “great worth” (Boro 2014: 142n381), emphasizing England’s positive qualities even at moments when the country is in trouble. In both of these passages, Tyler’s references to England’s wealth suggest that her readers should be proud to belong to such a country.

Tyler most obviously demonstrates her national pride with this declaration: “never England more flourished of knights, nor never nation was like to England” (1578/2014: 140). Boro notes the added “sense of patriotism” that Tyler brings to this moment with an emphasis on her country’s supremacy (Boro 2014: 140n372). Far from the country stripped of its knights by its prince’s disappearance, England has become a hub for great knights and knightly deeds. In fact, it surpasses all other nations, none of which can compare to it.

In addition to these compliments, Tyler refers to a specific figure representing England’s great tradition of knighthood: Sir Gawain. Tyler (1578/2014) describes one of the English knights, Brandidarte, as “a brave knight and as bold as Gawain” (141). Ortúñez’s text describes Brandidarte only as “one of the best knights…in Great Britain” (“vno de los mejores caualleros…en la gran Bretaña,” 1555/1617: 90). Tyler adds the specific reference to Gawain. This knight’s fame is established in medieval romances such as Sir Gawain and the Green Knight, where Gawain is “Of all knights on earth most honored” (line 914). Unfortunately, as with the case of Prince Edward and Trebatio, Tyler’s true hero is the Greek-Hungarian Rosicleer. Poor brave British Brandidarte is easily overthrown by the current giant. In a way, this defeat makes the reference to Gawain even more fitting, since in the later tales of Arthur’s court (such as those of Sir Thomas Malory), Gawain, formerly the greatest of the English knights, is less competent than the new-model French knight, Lancelot. Nonetheless, by comparing an English knight to the famous Sir Gawain, Tyler evokes a long tradition of excellent English knighthood, not to mention the great English romances that praise such knights. Tyler thus fully “Englishes” this scene by enhancing England’s prestige, drawing upon English literary and chivalric history, and evoking a strong sense of national pride.


In conclusion, the Mirror’s position as a harbinger of English interest in continental romance is no accident. By choosing a work with English characters and settings, and by applying the framework of translatio in such a way as to appeal to English readers’ cultural suppositions as well as their national pride, Margaret Tyler shapes the Spanish Espejo into a tale that captured the attention of English audiences. In addition, through her prefatory material, Tyler evokes cultural debates about women’s roles in writing and politics; her treatment of characters such as Briana and Olivia suggests Tyler’s willingness to re-evaluate the Espejo’s position on women in an English literary context. As critics grant more attention to translation as a creative practice (perhaps especially for women writers), Tyler’s skill in choosing the Espejo as well as “Englishing” it into the Mirror should be acknowledged as a particularly significant example of both literary and cultural translation.


Agorni, Mirella (1998) “The Voice of the ‘Translatress’: From Aphra Behn to Elizabeth Carter” The Yearbook of English Studies 28: 181-95.

Arcara, Stefania (2007) “Margaret Tyler’s The Mirrour of Knighthood: Or how a Renaissance translator became ‘the first English feminist” inTRAlinea, vol. 9 URL: http://www.intralinea.org/archive/article/1639. Accessed December 2017.

Bistué, Belén (2011) “The Task(s) of the Translator(s): Multiplicity as Problem in Renaissance European Thought” Comparative Literature Studies 48, no. 2: 139-64.

Boro, Joyce (ed.) (2014) Margaret Tyler’s Mirror of Princely Deeds and Knighthood, London, MHRA.

Budra, Paul (1992) “The Mirror for Magistrates and the Politics of Readership,” Studies in English Literature, 1500-1900 32, no. 1: 1-13.

Carrell, Jennifer Lee (1994) “A Pack of Lies in a Looking Glass: Lady Mary Wroth’s Urania and the Magic Mirror of Romance” Studies in English Literature, 1500-1900 34, no. 1: 79-107.

Carron, Jean-Claude (1988) “Imitation and Intertexuality in the Renaissance” New Literary History 19, no. 3: 565–79.

Coad, Katherine (1996) “Introductory Note” to The Mirrour of Princely deeds and Knighthood: Wherein is shewed the worthinesse of the Knight of the Sunne, and his brother Rosicleer, sonnes to the great Emperor Trebetio: with the strange love of the beautifull and excellent Princess Briana, and the valiant actes of other noble Princes and Knightes, by Margaret Tyler, in The Early Modern Englishwoman: A Facsimile Library of Essential Works, Part 1: Printed Writings, 1500-1640, Volume 8: Margaret Tyler, ed. Betty S. Travitsky and Patrick Cullen, Aldershot, Scolar Press: ix-xi.

Cooper, Helen (2004) The English Romance in Time: Transforming Motifs from Geoffrey of Monmouth to the Death of Shakespeare, Oxford, Oxford University Press.

Clarke, Danielle (2001) The Politics of Early Modern Women’s Writing New York: Routledge.

de Pizan, Christine (1521/2014) The Boke of the Cyte of Laydes, trans. Brian Anslay, ed. Hope Johnston, Tempe, AZ: ACMRS.

Elizabeth I (1544/2009), trans. Le Miroir de L’Âme Péchesse by Marguerite of Navarre, Elizabeth I and Her Age, ed. Donald Stump and Susan M. Felch, New York: Norton, 16.

Escobedo, Andrew (2004) Nationalism and Historical Loss in Renaissance England: Foxe, Dee, Spenser, Milton, Ithaca: Cornell University Press.

Ferguson, Margaret W. (1988) “A Room Not Their Own: Renaissance Women as Readers and Writers” The Comparative Perspective on Literature, ed. Clayton Koelb and Susan Noakes, Ithaca: Cornell University Press, 93-116.

Forrest, William (1558/1875) The History of Grisild the Second, ed. W.D. Macray, London, Chiswick Press.

Hackett, Helen (2000) Women and Romance Fiction in the English Renaissance, Cambridge, Cambridge University Press.

Herbert, Mary Sidney, trans. (1595) The Tragedy of Antonie by Robert Garnier, London: Ponsonby.

---, and Philip Sidney, trans. (2009) The Sidney Psalter ed. Hannibal Hamlin, et al., Oxford: Oxford University Press.

Hutchinson, Robert (2013) The Spanish Armada, New York, St. Martin’s.

“inheritrix, n.”, “inheritrice, n.”, and “inheritress, n” (2015) OED Online, Oxford University Press, March 2015 www.oed.com Accessed 22 May 2015.

Johnston, Hope (2014) “Introduction” to The Boke of the Cyte of Laydes by Christine Pizan, trans. Brian Anslay, ed. Hope Johnston, Tempe, AZ: ACMRS.

Krontiris, Tina (1988) “Breaking Barriers of Genre and Gender: Margaret Tyler’s Translation of The Mirrour of KnighthoodEnglish Literary Renaissance 18, no. 1: 19-39.

--- (1997). Oppositional Voices: Women as Writers and Translators of Literature in the English Renaissance. London: Routledge.

Lee, Christine S. (2014) “The Meanings of Romance: Rethinking Early Modern Fiction.” Modern Philology 112, no. 2: 287-311.

Locke, Anne Vaughan, trans. (1560) Sermons of John Calvin, vpon the songe that Ezechias made after he had been sicke, and afflicted by the hand of God, conteyned in the 38 Chapiter of Esay London: Day.

Lucas, Caroline (1989) Writing for Women: The Example of Woman as Reader in Elizabethan Romance Philadelphia: Open University Press.

Lucas, Scott (2003) “The Consolation of Tragedy: A Mirror for Magistrates and the Fall of the ‘Good Duke’ of Somerset,” Studies in Philology 100, no. 1: 44-70.

Magaw, Barbara Louise (1973) “The Female Characters in Prose Chivalric Romance in England 1475-1603,” PhD diss., University of Maryland.

Maly, Willy (2003) Nation, State and Empire in English Renaissance Literature: Shakespeare to Milton New York, Palgrave Macmillan.

Martin, Randall (ed.) (1997) Women Writers in Renaissance England, London. Longman.

McDermott, James (2005) England and the Spanish Armada: The Necessary Quarrel, New Haven, CT, Yale University Press.

Mentz, Steve (2006) Romance for Sale in Early Modern England: The Rise of Prose Fiction, Aldershot, Ashgate.

Nederman, Cary J. (2009) “Translatio imperii: Medieval and Modern,” in Lineages of European Political Thought: Explorations Along the Medieval/modern Divide from John of Salisbury to Hegel, Washington, D.C., Catholic University of America Press: 177–189.

Ortúñez de Calahorra, Diego (1555/1617) Espejo de Príncipes y Cavalleros, Juan Lanaja y Quartanet, Universidad de Carayoca. Google Books. Accessed 14 June 2017.

Roper, Margaret More, trans. (1531) A deuoute treatise vpon the Pater noster, made fyrst in latyn by the moost famous doctour mayster Erasmus Roterodamus, and tourned in to englisshe by a yong vertuous and well lerned gentylwoman of xix. yere of age. London: Berthelet, [1526]; 2nd ed. London: Berthelet, 1531.

Schleiner, Louise (1992) “Margaret Tyler, translator and waiting woman” English Language Notes 29, no. 3: 1-8.

--- (1994) Tudor and Stuart Women Writers, Bloomington, Indiana University Press.

“schoolpoint, n.” OED Online, Oxford University Press, March 2015, www.oed.com Accessed 22 May 2015.

Schwyzer, Philip (2004) Literature, Nationalism, and Memory in Early Modern England and Wales, Cambridge: Cambridge University Press.

Sir Gawain and the Green Knight (2001) Sir Gawain and the Green Knight; Patience; Pearl: Verse Translations, trans. Marie Borroff, New York, Norton, 1-74.

Snook, Edith (2005) Women, Reading, and the Cultural Politics of Early Modern England Aldershot: Ashgate.

Spenser, Edmund (1596/2007) The Faerie Queene, ed. A.C. Hamilton, Harlow, Pearson Longman.

Summit, Jennifer (2001) Lost Property: The Woman Writer and English Literary History, 1380-1589, Chicago and London: University of Chicago Press.

Swift, Helen J. (2008) Gender, Writing, and Performance: Men Defending Women in Late Medieval France, 1440-1538, Oxford: Clarendon Press.

Tyler, Margaret (1578/2014) Margaret Tyler’s Mirror of Princely Deeds and Knighthood, ed. Joyce Boro, London, MHRA.

Uman, Deborah (2012) Women as Translators in Early Modern England, Newark, University of Delaware Press.

---, and Belén Bistué (2007) “Translation as Collaborative Authorship: Margaret Tyler’s The Mirrour of Princely Deeds and Knighthood,” Comparative Literature Studies 44, no. 3: 298-323.

Winston, Jessica (2004) “A Mirror for Magistrates and Public Political Discourse in Elizabethan England,” Studies in Philology 101, no. 4: 381-400.


[1] On the “mirror” genre, see Winston (2004), Lucas (2003), Budra (1992), and Uman and Bistué (2007: 315, 318).

[2] East, Tyler’s publisher, printed second and third editions of the Mirror as well as five of the nine volumes of Ortúñez’s Espejo (Boro 2014:4). None of these other translations were by Tyler.

[3] See Nederman (2009) and Carron (1988).

[4] On translation: Bistué (2011), Uman (2012), Uman and Bistué (2007), Agorni (1998), Arcara (2014). On romance: Carrell (1994), Hackett (2000: 159-93), Lee (2014: 287-311), Magaw (1973).

[5] For the preface, see Agorni (1998: 182-83); Arcara, (2014); Martin, Women Writers in Renaissance England, 18-24. Mentz (2006: 125) mentions Tyler’s preface as evidence that “there were some women who read fiction,” which, though true, seems too simple for this complex rhetorical piece. See also Krontiris (1988), Schleiner (1992) and (1994).

[6] See Hutchinson (2013) and McDermott (2005).

[7] See also Schleiner (1994: 18-23), Ferguson (1988: 95-116), Lucus (1989), and Snook (2005).

[8] See Cooper (2004: 6) on the “educational” aspect of romance.

[9] This list is drawn from Arcara (2014). 

[10] Tyler’s most famous declaration in her defense is the following: “whatsoever the truth is—whether that women may not at all discourse in learning for men lay in their claim to be sole possessioners of knowledge, or whether they may in some manner that is by limitation or appointment in some kind of learning—my persuasion hath been thus: that it is all one for a woman to pen a story as for a man to address his story to a woman” (1578/2014: 50).

[11] Schleiner (1994) notes this tendency in Tyler’s preface: “She has not after all gone so far as to write an original piece” (253n21). Krontiris (1997: 48) also mentions “the difficulty in speaking out openly;” in particular, Tyler “does not specify the aspects that make it [the Mirror] controversial.” Krontiris further describes Tyler’s refusal to “expurgate” her text, particularly with regards to women’s sexuality, as one such controversial aspect (1997: 62).

[12] Hutchinson (2013: xii) presents the Spanish Armada as “the climax to a war of religion, the Catholic Church versus the fledgling Protestant state of England.” Catherine of Aragon herself was a figure of much consideration in the sixteenth century; on the opposing side to those who saw Catherine as a threatening Catholic figure, William Forrest’s 1558 poem The History of Grisild the Second compares Catherine to the proverbially patient Griselde.

[13] Although Tyler’s first audience was mostly Protestant, scholars are still plagued by the question of whether Tyler’s Mirror itself reveals Protestant or Catholic sympathies. Tyler’s association with the Howard family as well as her knowledge of Spanish are often taken as signs of Catholicism (Boro 2014:6, Martin 1997: 16). Boro (2014: 9) further argues that the Mirror’s publisher, Thomas East, had “personal and professional connections within the Catholic community” that may have led him to print the Mirror. One of Tyler’s later positions, however, was in the household of Nathaniel Bacon, a Protestant (see Schleiner 1992: 3-4). Religious references in the Mirror itself are similarly inconclusive. For instance, Tyler omits a character’s habit of hearing daily mass, but she adds a description of Briana living like an “anchoress” and includes fasting in Briana’s rituals (Boro 2014: 231n664, 72n100, 73n106). The omissions may reflect Tyler’s desire to avoid antagonizing a primarily Protestant audience, but her references to anchoresses and to fasting may have the opposite effect.  Certainly, in comparison to another sixteenth-century romance, Spenser’s Faerie Queene, Tyler’s Mirror is positively kind to Catholic characters. Tyler thus remains relatively neutral on questions of religion even as she makes no secret of the Mirror’s Spanish origins, with all the political implications inherent in her choice of work.

[14] Despite this reference to hospitable translation, however, although Tyler frequently references the Mirror’s foreign origins, not once does she mention Ortúñez’s name. Tyler’s own name is prominent in her paratext; her dedication is signed “Margaret Tyler,” while both the title page and the epistle “To the Reader” indicate Tyler’s authorship with the initials “M.T.” By including these markers of her identity, Tyler erases the other name in the equation: that of Diego Ortúñez de Calahorra, author of the Espejo. Tyler reduces the Spanish author to a shadowy background figure, a nameless “stranger” whose work Tyler both mediates and authorizes by presenting it to an English readership. When gender enters the equation, this language of appropriation becomes more fraught—particularly when the writer is a woman and the translator a man, as is the case with Christine de Pizan, whose Boke of the Cyte of Ladyes was translated into English by Brian Anslay in 1521. In situations like these, scholars, most notably Jennifer Summit (2001: 61-71), see an appropriation of power whereby a man co-opts a woman’s words and in some sense her identity. Others, such as Swift (2008: 184) and Johnston (2014: xliv-li) argue that automatically reading translation as appropriation simplifies what are in fact a wide range of individual and cultural factors.

[15] Although see Florio’s preface to Montaigne’s Essays, which describes in similar terms how he “transported it from France to England; put it in English clothes; taught it to talke our tongue.”

[16] In some ways, Tyler’s adaptations of female characters create a challenge to other writers, one accepted most famously by Mary Wroth in her rich Urania (1621).

[17] Other female characters have similarly conventional roles; Krontiris (1988) notes that women are most likely to be the victims in need of rescue by knights, from Calinda, whom Rosicleer saves from a giant (136), to the Duchess Elisandra, whom the Knight of the Sun rescues from false charges of adultery (230).  Tyler’s substitution of Briana for Claradiana in her title, then, fits with this trend of conventional female characters.

[18] Boro (2014: 18-25); Uman and Bistué (2007: 300) particularly note Tyler’s “colloquialisms” and “alliteration.”

[19] Tyler (1578/2014) does not, however, change every piece of Spanish vocabulary or cadence. For instance, she usually refers to Britain as “the Great Britain” (see 142 for an example), which seems to be a holdover from the Spanish “la gran Bretana” rather than common English usage (see Ortúñez 1555/1975: 33). Similarly, she retains the Spanish “Oliverio” as the King of England’s name instead of the more English “Oliver.”

[20] Boro (2014: 23) confirms that this phrase is Tyler’s addition.

[21] The OED uses sample quotations from Roger Ascham (1568) and Sir Philip Sidney (1587), demonstrating that this usage was current in Tyler’s day.

[22] Tyler also uses the phrase “school points” when she describes Rosicleer as an anxious scholar composing his letter to Olivia: “For the better understanding whereof, you must imagine a young scholar but lately entered into school points overseeing of his theme before he bring it to the review of his schoolmaster. And believe me, in far greater doubt hung Rosicleer of his lady’s liking than the boy doth of his master’s” (Tyler 1578/2014: 161).

[23] Boro (2014: 22, 23) ultimately portrays this scene as a failed attempt by Tyler to anglicize the text, since “Trebatio’s gruesome actions” in the “shocking scene” of Edward’s defeat produce conflicting emotions without a moral center.

About the author(s)

Rachel M. De Smith Roberts is Assistant Professor of English at North Greenville University, where she teaches writing and literature. She holds a B.A. in English from Dordt College, an M.A. in English from Creighton University, and a PhD in English from Baylor University. From 2011 to 2014, she was a member of the Lilly Graduate Fellows Program (Fourth Cohort). Her scholarship focuses on early modern women writers, particularly four women who wrote in male-dominated genres: Margaret Tyler, Anne Dowriche, Mary Wroth, and Elizabeth Cary.

Email: [please login or register to view author's email address]

©inTRAlinea & Rachel Roberts (2017).
"Englishing A Spanish Romance Translating Spanish Rivalry into English Patriotism in Margaret Tyler’s Mirror of Princely Deeds and Knighthood"
inTRAlinea Volumes
Edited by: {specials_editors_volumes}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2264

Interpreters Working with Children in Italy

Profile, Role and Expectations

By Amalia Amato & Gabriele Mack (University of Bologna, Italy)

Abstract & Keywords

This paper deals with interpreting for children involved in criminal investigations in Italy. Within the CO-MINOR IN/QUEST research project (2013-2014), in order to investigate this subject a web-based questionnaire was submitted to justice professionals and police officials, psychologists, interpreters and child support workers all over Europe. Data collected from experienced Italian respondents to the questionnaire are presented to illustrate the status and (self)perception of interpreters working with children in this field in Italy.

Keywords: interpreting, children, vulnerability, criminal investigations, survey, roles, interpreter training

©inTRAlinea & Amalia Amato & Gabriele Mack (2017).
"Interpreters Working with Children in Italy Profile, Role and Expectations"
inTRAlinea Volumes
Edited by: {specials_editors_volumes}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2263

1. Introduction

This paper deals with interpreting for children involved in criminal investigations in Italy.[1] In order to explore this subject a web-based questionnaire was submitted to justice professionals, law enforcement officials, psychologists, interpreters and child support workers all over Europe in the framework of CO-MINOR IN/QUEST, a EU-funded research project (2013-2014).

Data collected from Italian respondents to the questionnaire is presented and discussed here to illustrate the status and (self)perception of interpreters working with children in Italy. The purpose of this paper is to discuss the answers given by interpreters about their profile, their role, their main (perceived) challenges and expectations, and their suggestions for improvement. In some cases the interpreters’ responses are compared with those given by other professionals to highlight agreement or disagreement on specific issues. In conclusion some of the indications given by respondents will be proposed as possible future actions to be undertaken in this field in order to comply with the internationally recognised principle of the “best interest of the child” (UN Convention on the Rights of the Child, 1989).

The content of this article was first presented at the Critical Link 8 conference “Critical LinkS – a New Generation: Future-proofing Interpreting & Translating” held on 29 June – 1 July 2016 at Heriot-Watt University in Edinburgh (Scotland). Last update of statistical information provided in this paper was October 2016.

2. Why interpreting for children deserves special attention

There are several reasons to embark on research work on interpreting for children involved in criminal proceedings. First and foremost, there is a growing awareness that children[2] are vulnerable subjects (often in many ways) and need special safeguards and care. A second equally important reason is the growing number of migrant children arriving in Europe, who, moreover, are often unaccompanied. Interpreting for children is, therefore, a growing and challenging field of activity for today’s and tomorrow’s professionals, who must learn how to deal with new scenarios. Talking to children is different from communicating with adults. This can be seen in particular with infants, but it also holds true for older children (see, for example, Owens 1984/2015; Lefevere 2010; Winter 2010). And language development is an important aspect that comes into play when communicating with a child who does not speak the language of the local institutions in a legal context, especially during criminal proceedings.

The growing attention paid to this matter by European lawmakers is evident, starting with the EU Charter of Fundamental Rights (2000) under Articles 24 (The rights of the child), 20 (Equality before the law) and 47 (Fair trial), and in subsequent actions within the framework of the Stockholm Programme[3]. So far this programme has produced three directives directly relating to the rights of children: Directive 2011/36/EU on prevention and combating trafficking in human beings and protecting its victims; Directive 2012/29/EU establishing minimum standards on the rights, support and protection of victims of crime and, more recently, Directive 2016/800 on procedural safeguards for children who are suspected or accused persons in criminal proceedings. Directive 2016/800 in Whereas (48) acknowledges the “inherent vulnerability” of children and the fact that there may be forms of discrimination on the ground of language (65). It also recognises communication difficulties based on special needs a child may have (55). Article (20) is devoted to specific training for staff of law enforcement authorities and of detention facilities who handle cases involving children (1.), judges and prosecutors (2.), lawyers (3.) and those providing children with support and restorative justice services. This approach was already enshrined in Article 6 of Directive 2010/64 on the right to interpretation and translation in criminal proceedings, the very first EU directive in the field of justice. One noteworthy aspect is that none of these directives mentions training for interpreters, leaving the question open as to whether children’s rights can be effectively safeguarded when not all the communication partners are duly qualified.

2.1 Vulnerability

The European Commission Recommendation of 27 November 2013 on procedural safeguards for vulnerable persons suspected or accused in criminal proceedings (2013/C 378/02), instead of defining the concept of vulnerability, names four characteristics that can be related to “vulnerable persons”: age, mental or physical condition and disabilities (Preamble, Whereas 1). Previously, in the Green Paper on Procedural Safeguards for Suspects and Defendants in Criminal Proceedings throughout the EU (2003), the Commission had mentioned a number of potentially vulnerable categories, giving a non-exhaustive list of eight groups: foreign nationals, especially but not solely those who do not speak the language of the proceedings; children under the UN Convention on the Rights of the Child (Article 40); persons who are vulnerable as a result of their mental or emotional state; persons who are vulnerable as a result of their physical state; persons who are vulnerable by virtue of having children or dependants; persons who cannot read or write; persons with refugee status under the 1951 Refugee Convention or other beneficiaries of international protection and asylum seekers; persons dependent on alcohol or drugs (Green Paper 2003, 32-34). The European Court of Human Rights in its judgements has identified an even broader number of specific groups that are vulnerable because they have suffered considerable discrimination, namely on account of their gender, sexual orientation, race or ethnicity, mental faculties, physical disability, HIV/AIDS condition, and status of asylum seeker (Justicia European Rights Network 2013). Children, especially migrants, may fall into most of the categories mentioned above and are, thus, often subject to multi-layered, multi-factorial vulnerability[4].

3. Facts and figures about non-Italian speaking children in Italy

A look at the statistics shows the scale of the phenomenon. According to a European Commission Staff Working Document (2013), “The number of children facing criminal justice is approximately 1,086,000 across the EU, i.e. 12% of the European population facing criminal justice each year” (2013: 4). “About 1.6% of the total prison population within the EU are children. (…) These figures are unlikely to decrease by any significant amount in the coming years” (2013: 30), and non-nationals are a significant part of these children.

In September 2016, according to data published by the Department for Juvenile Justice and Alternative Measures of the Italian Ministry of Justice, 4,917 non-Italian minors (with an increasing trend since 2007; see Figure 1) were entrusted to the care of the Department’s social services, 1,407 of whom appeared in the statistics for the first time that year, with 1,205 of them under 16 years of age when they were included in the statistics for the first time (Dipartimento Giustizia Minorile e di Comunità 2016: 5). The regions of origin were roughly equally divided between Africa (35%; mainly Morocco), EU-member states (27%; mainly Romania) and other European countries (26%; mainly Albania) (ibidem: 11).


Figure 1: Minors taken care of by Childhood Social Services of the Italian Ministry of Justice between 2007 and 2015 according to nationality: Italians in blue - 15,913 in 2015, foreigners in red - 4,625 in 2015; the total figure for 2015 is 4,625
(Dipartimento Giustizia Minorile e di Comunità 2016: 9).

These figures do not distinguish between children living with their parents and the steadily growing number of so called ‘unaccompanied alien minors’. Under Italian law, this term covers all non-Italian or non-EU-nationals who, without having applied for asylum, live in the country without being assisted and represented by their parents or by another legal representative (Giovannetti 2016: 22). According to a report published by the Association of Italian Municipalities (ANCI), after a slight decrease in 2009 and 2010, the number of non-Italian separated children reached 13,523 in 2014 (Giovannetti 2016: 45; Figure 2):


Figure 2: The number of non-Italian separated children in touch with or taken care of by Italian social services between 2006 and 2014 (Giovannetti 2016: 45).

Obviously the need for children to communicate through an interpreter with legal institutions using a language they are unable to understand and/or speak greatly depends on how long they have been living in the country. Following recent events in the Middle East and along the southern shores of the Mediterranean, the number of migrants arriving in Italy is soaring. Many of these migrants are children, both accompanied and unaccompanied, and consequently there is a growing need to provide them with interpreting services. These children may need interpreting in emergency situations after fleeing from their country of origin, but they also need interpreting when they are suspected or accused of having committed criminal offences, when they are victims or witnesses (for example of abuse, sexual or economic exploitation, trafficking of human beings or other crimes) and when they are involved in refugee status proceedings, international protection or asylum hearings (although asylum hearings in Italy are classified as administrative proceedings).

4. Interpreting for children: a new research focus

Although interpreting for children only became a separate focus of research quite recently, there is considerable interest in this topic with many publications. Rather than providing an exhaustive overview of the international literature, which is beyond the scope of this paper, a glimpse is provided of the main areas of investigation before proceeding with the CO-MINOR IN/QUEST survey.

4.1 Main areas of investigation

Research on educational interpreting for deaf and hard of hearing children started in the 1990s, also thanks to institutions, such as the Laurent Clerc National Deaf Education Center at Gallaudet University and Boys Town National Research Hospital, and researchers such as Brenda Schick at the University of Colorado-Boulder (http://www.classroominterpreting.org/). The level of awareness of the complexity of this kind of interpreting task and the need for collaboration among all professionals involved is very high among these institutions and researchers, and is reflected in a remarkable number of publications (for an overview see Smith 2015). A second traditional field of study since the late 20th century has been interpreting in pediatric settings (for an overview see Leanza and Rocque 2015), both in North America (Kirmayer et al. 2001; Alvarado-Little 2004; Phoenix Children's Hospital 2008) and in Northern Europe (Granhagen et al. 2016; Together for Short Lives, n.d.). A growing field of interest since the 1990s has also been interpreting for children and young adults in mental health settings in Canada (Rousseau et al. 2011), the UK by a British research group (Raval 1996, 2005; Loshak 2003), and later Sweden (Jarkman 2005, 2013), Germany (Quindeau and Rauwald 2016) and South Africa (Penn and Watermeyer 2014). Migration-related traumas are frequently mentioned in this context. Detailed research has also been carried out in the field of child protection (for example, Chand 2005; Hegemann 2009; Kriz and Skivenes 2010; Norström et al. 2011; Nilsen 2015), which has been made use of in user guidelines (for example, London Safeguarding Children Board 2016). The need to assist refugee children led to the concept of “migrationssensibler Kinderschutz” (migration-sensitive child protection) developed by Teke (2016).

Last but not least, interpreting for children in legal settings is the object of a growing number of publications all over the world – from South Africa (Matthias and Zaal 2002) to the USA (Hiltz and Anderson 2003), Sweden (Keselman 2009; Keselman et al. 2010; Linell and Keselman 2010), and the UK (Pollitt and Little 2003), once again with a keen interest in migration-related aspects, such as asylum hearings.

One of the most ambitious research projects on interpreting for children so far is CO-Minor-IN/QUEST – Cooperation in interpreter-mediated questioning of minors (2013-2014) funded by the EU Commission, which also produced the first book on the topic (Balogh and Salaets 2015). The data discussed in the central part of this paper were collected as part of this research project. Other outputs of this research are Van Schoor (2013); Böser et al. (2014); Salaets and Balogh (forthcoming).

For Italy, it is extremely difficult to find reliable, up-to-date information on the language and interpreting needs of children, let alone comprehensive studies and literature on best practices. In most official documents attention to problems related to communication, language and the need for interpreting is very scarce, when not totally absent. According to the already quoted ANCI report, in 2013-14 “interviews with the minor” were the most frequently adopted measure to safeguard unaccompanied alien children, followed by “putting them in a safe place” (Giovannetti 2016: 218-19). Problems with the activities of cultural mediators acting as interpreters in most migration contexts are presented in several studies about separated children, stating for example that “for crucial figures such as cultural mediators and ethno-psychiatrists there is no drive towards stabilisation within the service, not even towards establishing professional registers guaranteeing the qualifications required” (NET FOR U 2014: 14; our translation). Yet, understanding the alternatives offered by local services during the very first contacts is crucial for foreign children. This aspect emerges very vividly in interviews conducted within the research project PUCAFREU – Promoting unaccompanied children's access to their fundamental rights in the EU (Rozzi 2013: 63, 67). One of the conclusions of the study is that action for children carried out by mixed teams “including also cultural mediators, non-Italian educators and peer educators” proved to be extremely effective in achieving a number of goals: making contact with newly arrived children as soon as possible, informing them about their rights, possible steps towards integration and services available, creating a bond of trust and assisting these children to make informed choices (ibidem: 82).

4.2 The CO-MINOR IN/QUEST survey

Extensive research work on interpreting for children in legal settings was conducted between October and December 2013 by the CO-MINOR IN/QUEST research team. A survey was carried out via the Qualtrics Online Survey Programme, a very flexible tool which made it possible to reach professionals in different European countries in a fast, simple way. An online questionnaire, drafted in English and then translated into the languages of the participant countries (French, Dutch, Hungarian and Italian), was advertised by each partner among justice professionals and policing officials, psychologists, interpreters and child support workers, but also respondents from non-partner countries were invited to participate on the basis of personal contacts. The choice of the tool and of the distribution method proved to be successful. The questionnaire was accessed by over 1,000 potential respondents and was filled in (at least in its main parts) by 848 respondents from 16 countries (mainly Belgium, France, Hungary, Italy, the Netherlands and the UK)[5]. The aim of the questionnaire was to provide a picture of the current situation as regards interpreting for children in the countries participating in the project. It had both closed (check boxes) and open questions for additional comments[6]. This research design made it possible to analyse the data both quantitatively and qualitatively.

The data collected from the online survey was configured directly to the Customer Relationship Management System of Qualtrics, which makes it possible to maintain a high level of data integrity and reduce errors. This programme automatically or semi-automatically identifies and, when possible, corrects errors in large data sets.

The questionnaire was designed on the basis of a literature review and a workshop involving three legal practitioners (a juvenile judge, a juvenile lawyer and a trainer in forensic interview techniques), two psychologists (one of them a forensic psychologist) and five interpreters (including one sign language interpreter). The contributions made by the different professionals were very useful in identifying issues stemming from first-hand experience in the field of questioning children that needed to be investigated. The whole questionnaire preparation process was supervised by an expert from the Eszter Foundation – a criminologist, consultant and research expert for the Hungarian National Committee of UNICEF with considerable experience in quantitative and qualitative research.

The questionnaire had an introduction explaining the purpose of the project and, after choosing their preferred language, respondents were immediately asked to select the country and region where they worked and one category out of interpreting, justice and policing, psychology and other. In the subsequent analysis the categories were reclassified into five different categories: justice and policing, interpreting, psychology, child support and other. Respondents were then asked to give their job title, their experience in working with children and the age groups of children they worked with, the frequency and type of cases (crimes) they had experience of. All respondents, apart from the interpreters, were asked whether they had received training to work with interpreters and interpreters had to indicate what type of training they had received. All respondents were asked whether they had received training to work with children and again it was possible to specify the length of training and whether it was formal training or training on the job. The next question was about experience of interpreter-mediated encounters with children. Those who answered ‘yes’ could then proceed to the second and afterwards the final part of the questionnaire. Those who answered ‘no’ were redirected to the final demographic section of the questionnaire. This section was meant to gather information about age, gender, highest level of education/qualifications, mother tongue and language combination (for interpreters) and membership of a professional organisation.

The second part of the questionnaire asked questions about the main challenges perceived by the respondent in relation to an interview with a child (in the case of interpreters) or an interview mediated by an interpreter (in the case of other professionals). The next two questions focused on “before the interview”, namely whether a briefing was provided for the interpreter and how the briefing was provided. Then the questionnaire had a part devoted to what happens “during” the interview. Respondents were asked to choose one or more answers about what is required from an interpreter when working with children. Immediately after, ten statements about the function of the interpreter were presented and respondents had to express their agreement or disagreement using a 5-point-Lickert scale. The following section was entirely devoted to the seating arrangement and asked respondents to express their preference on the basis of drawings. The last part of the questionnaire investigated what happens after an interpreter-mediated child interview takes place: is there a de-briefing and if so, how is it carried out. The last question before the demographic section was about what respondents felt should be improved in interpreter-mediated interviews with children, a totally open question.

The following section contains a detailed analysis of the answers given by Italian respondents about their perception of the interpreter’s role, their opinions about the main challenges of interpreting for children and the special skills that are deemed necessary to carry out this activity successfully.

4.3 Italian respondents

In the survey, 117 Italian respondents out of 230 explicitly stated they had experience in working with children (and with interpreters, in the case of other professionals). Among these experienced respondents, 32 were interpreters (including 4 sign language interpreters), 85 other professionals (19 justice and policing professionals, 4 psychologists and 62 child care social workers). For the purpose of this paper all non-interpreters will be considered as a single group and referred to as “other professionals”. Another caveat worth mentioning for correct data reading and interpretation is that since not all respondents answered all the questions and some questions allowed multiple answers, the figures about responses discussed below can vary from case to case and the sum of the percentages sometimes exceeds 100%.

5. Focus on interpreters: profile

The first aspect to discuss is the profile of the interpreters who work with children that emerges from answers to the questions concerning gender, age, education, training in specific interpreting techniques and in working with children, training in legal interpreting, age of children they usually work with, language combinations, years of experience with children involved in criminal proceedings and type of crimes.

On the basis of the answers provided by the 32 interpreters who responded to our questionnaire it is possible to draw an average profile of an interpreter who works with children involved in criminal proceedings in Italy. The interpreter is generally female (92% of our sample), aged 41-54 (out of 26 respondents, 73% were aged 41-54, nobody under 30 and only 4 over 54 years of age) with a high level of education (almost 70% of respondents had a Bachelor’s degree and over 50% a Master’s degree). In our sample 21 (81%) were Italian native speakers, two were Arab native speakers, together with one Italian sign language, one French, one Russian and one Ukrainian native speakers. The language combinations of the interpreters in our sample mainly included Western European languages: English, French, German and Spanish, followed by Arabic, Chinese, Russian, Ukrainian, Romanian and Italian Sign Language.

As far as training in specific interpreting techniques and subjects is concerned the highest percentage (60%) had received training in sight translation; less than half of our sample stated that they had received training in legal interpreting and in consecutive interpreting. Another question enquired about specific education or training to work with children. All respondents answered this question, but only 3 out of 32 had received some specific training in working with children.

When asked about their experience with children, 7% of respondents said they had worked with children less than one year, 77% said they had 10 years or more experience and 16% had 4-9 years of experience. The following question concerned the number of encounters with children interpreters had been hired for in the previous 3 years: 46% of respondents said between 1-4 encounters, 27% said over 40 and 27% 5-20 encounters. As for the age groups of the children they had worked with, 52% declared they had experience with children aged 11-14, 96% with children aged 15-18, 30% with children aged 7-10 years and 3% with children aged 0-3.

In terms of crime typology, the two major offences were sexual offences and child abuse or neglect, while the category ‘Other’ was mainly related to migration issues.

5.1 Functions of the interpreter when interpreting for children

Respondents were presented with ten statements about the interpreter’s function and asked to express their degree of agreement or disagreement on a 5-point Lickert scale. They had to express their degree of agreement/disagreement on two statements about support, two statements on expectations about translation, five statements on different types of possible initiatives by the interpreter, and finally whether the interpreter should give an opinion or not. The respondents about the functions of interpreters were 27 interpreters and 80 other professionals, with the answers of the two groups compared. The answers were grouped together in three broader categories: agree, neither agree nor disagree, and disagree – a simplification compared to the original scale with the aim of making the information more easily accessible. Respondents also had the opportunity to comment on each statement. One general aspect emerging from these comments was that “neither agree nor disagree (at least for interpreters) often meant “it depends on the situation”.

The first two statements were about supporting one party during the interview: either the child or the interviewer. Concerning Statement 1: “The interpreter supports the minor through his/her interpretation and initiative” (Figure 3) most respondents of both groups expressed disagreement though in different percentages: other professionals (91%) were far more opposed to this statement than interpreters (67%).


Figure 3: Opinions of interpreters and other professionals about the possibility of the interpreter supporting the child during the interview.

When it comes to supporting the interviewer - Statement 2: “The interpreter supports the interviewer’s purposes through his/her interpretation and initiative” (Figure 4) the situation changes. Most interpreters still disagreed (63%), but only 36% of other professionals disagreed, while 58% of them would like the interpreter to support the interviewer's purposes.


Figure 4: Opinions of interpreters and other professionals about the possibility of the interpreter supporting the interviewer during the interview.

The second group of statements was about interpreting literally and interpreting faithfully, starting with Statement 3: “The interpreter interprets literally” (Figure 5). Here again there was some disagreement between the two groups, since the majority of other professionals (63%) wanted the interpreter to interpret literally, while interpreters mainly did not want to interpret literally (41%).


Figure 5: Opinions of interpreters and other professionals about literal interpretation.

When it comes to faithfulness - Statement 4: “The interpreter interprets faithfully” (Figure 6) - both groups (96% of interpreters and 85% of other professionals) largely agreed that an interpreter should interpret faithfully. This suggests that other professionals tend to identify literality in interpreting with faithfulness, while interpreters do so much less.

Interestingly, one professional (belonging to the justice and policing professions) commented that she did not understand the distinction made between literal and faithful, and concluded that “translation should be faithful to what the child wants to express”.


Figure 6: Opinions of interpreters and other professionals about faithful interpretation.

The following group of five Statements (5 to 9) was about the interpreter taking initiatives, in particular:

  • to explain social-cultural differences (Figure 7)
  • to explain technical terminology (Figure 8)
  • to adjust the language to the level of the minor (Figure 9)
  • to put the minor at ease (Figure 10)
  • to keep the communication flowing (Figure 11).

In this case the opinions were quite clear-cut and very much in line for all five statements: between 61% and 89% of respondents in both groups agreed that the interpreter should indeed take all these initiatives (and, therefore, also responsibility for it) while only between 0% and 15% disagreed.


Figure 7: Opinions of interpreters and other professionals about the interpreter’s initiative to explain social-cultural differences.


Figure 8: Opinions of interpreters and other professionals about the interpreter’s initiative to explain technical terminology.


Figure 9: Opinions of interpreters and other professionals about the interpreter’s initiative to use a child-friendly language.


Figure 10: Opinions of interpreters and other professionals about the interpreter’s initiative to put the child at ease.


Figure 11: Opinions of interpreters and other professionals about the interpreter’s initiative to keep the communication flowing.

In other words, other professionals expect interpreters to undertake many actions that might instead be considered part of their tasks. Explaining technical terminology or adjusting the language to the cognitive and language ability of the child are a case in point, with Article 20.1 of Directive 2016/800 identifying this as a responsibility of the staff of law enforcement authorities. Interestingly, however, the answers show that interpreters too feel obliged to accomplish all of these tasks. In all cases but one (technical terminology; Figure 8) they feel even more obliged than the other professionals. This is certainly a point that deserves further reflection and investigation.

As regards Statement 10: “The interpreter gives his/her opinion on the case” (Figure 12) both groups expressed almost the same percentage of disagreement (63% and 61% respectively), thus showing some sensitivity to the issue of neutrality. Nevertheless, almost 20% of other professionals would like to have the interpreter’s opinion, and 30% of interpreters do not rule out the possibility of giving their opinion, at least in principle.


Figure 12: Opinions of interpreters and other professionals about the interpreter expressing his/her opinion on the case.

5.2 Main challenges in interpreter-mediated interviews with children

Interviewing a child who does not speak the language of the investigation or proceedings is usually a multi-party communication event. Participants generally include at least the interviewer, who may be a representative of the law enforcement agency or a juvenile judge, a psychologist, a social worker, an interpreter and the child. It is, therefore, a complex interaction because of the different professionals involved, their perception about their own role and the role of the other participants, the dynamics of rapport building, working in two different languages through an interpreter and, last but not least, the issue of trust.

Some of these aspects were presented to respondents as potential challenges and they were asked to choose one or more among the following items: 1. Working with other professionals, 2. Working with the child, 3. Handling the interpreting process, 4. Other.

According to the data collected, the interpreters in our sample considered working with other professionals and working with the child as rather problematic (which may be well expected), but on the whole these two aspects ranked only second and third (with 41% and 48% respectively). What instead was seen as the main challenge, is handling the interpreting process, mentioned by 51% of the 27 interpreters who responded to this question. Since they are professional interpreters with experience in this domain, this could be related to the specificity of this particular situation; but, on the basis of the answers about the interpreter’s functions mentioned above, this result might also be put in relation to the number of initiatives the interpreter is supposed to take, both in her/his eyes and in those of the other professionals. A closer look at the interpreters’ specifications about their choice confirms this interpretation of the data.

Two interpreters’ comments are particularly enlightening about the interpreters’ perception of the challenges when working with children:

“Sometimes it is necessary for the interpreter to simplify concepts and language”

“It is always very difficult to combine language mediation and the need for psychological sensitivity”

The first comment reflects the idea that the interpreter feels responsible for adapting other professionals’ language to the child’s linguistic and cognitive developmental stage.

The second comment highlights psychological concerns, an issue that is recurrent in comments on several points of the questionnaire and, indeed, appears to be a specific difficulty in interviews with children.

Other professionals were asked to select what they considered the main challenges when working with an interpreter. They could pick one or more of the following options: 1. Interpreter’s poor understanding of interview techniques, 2. Interpreter’s lack of neutrality, 3. Interpreter’s insufficient cooperation, 4. Other.

Other professionals who responded to the question about the main challenge(s) concerning interpreting when interacting with a child in legal settings were 83. They mentioned first of all the interpreters’ poor understanding of interview techniques (52%) and lack of neutrality (33%), but under the category ‘other’ (39%) they added many other challenges they specified in 23 comments. The most frequently mentioned aspects pointed mainly to interpreters’ lack of skills or knowledge, in particular: insufficient knowledge of the criminal justice system, lack of psychological knowledge (both child psychology and handling emotional aspects of the situation) and insufficient accuracy in translation. The last point can be related to the high number of respondents who identified lack of neutrality on the part of the interpreter as a main challenge. Since other professionals do not trust the interpreter to be neutral, they believe that this is reflected in the interpreting performance. As a consequence, one might expect that trust in interpreters also suffers or at least is not that solid as it should be in a group of people who supposedly work together to protect the best interest of the child.

On the whole, all these comments point to a strongly felt need for training for both groups or, better still, joint training.

5.3 What other professionals require from interpreters

One question was addressed only to other professionals and was aimed at enquiring about special characteristics an interpreter should have in order to work with children. Respondents could select one or more characteristics they want interpreters to have among the following: 1. Same gender as the minor, 2. Previous experience with minors, 3. Previous experience with the same case, 4. None, 5. Other.

By far the most frequent answer was experience, mentioned by 85% of the 80 respondents. Once again, however, under the category ‘other’ many other requirements were mentioned, and they partly overlap with the challenges already discussed: knowledge of the criminal justice system, knowledge of psychology, (inter)cultural knowledge, relational skills, and ability to work in a team. It is not known whether the interpreters and other professionals who responded to the questionnaire work together regularly and for this reason the complaints expressed by other professionals cannot be directly referred to our interpreters’ sample. But as mentioned in section 5 above, less than 10% of interpreters in our sample had some specific training in working with children, and less than 50% of them declared they were trained in legal interpreting and in consecutive interpreting. There is definitely a lack of specialised skills that are necessary in order to work in this field. So once more the need for specific training for interpreters emerges from the survey. On the other hand, other professionals should not just point their fingers at interpreters. If they want them to develop relational skills and be able to participate in teamwork, as stated in the questionnaire, they should share their knowledge about interviewing techniques with interpreters, encourage and become involved in joint training, be aware of role boundaries and not expect the interpreter to be a ‘jack of all trades’. They should also be prepared to recognise the professional status of interpreters, be ready to place them on an equal footing with other professionals, and try to gain a better understanding of the complex task of interpreting. This is well reflected in the following comment by one interpreter who expressed her frustration about the lack of understanding among other professionals of what interpreting is about: “Police officers are not always cooperative towards interpreters. After all, they only have to translate…”.

5.4 Suggestions for improvement

One of the last questions for both interpreters and other professionals was: “What do you personally think is needed to improve interpreter-mediated encounters with minors?” The aspects most frequently mentioned by interpreters included: having a briefing with other professionals before the interview, having the opportunity to meet the child before the interview, receive support from other professionals, and having training for all parties involved. The responses provided by other professionals partly overlap: having a briefing with the interpreter, need for experienced interpreters, need to work with the same interpreter in order to have continuity and be able to build teamwork and, last but not least, training for all parties involved. These items are not listed exactly in their order of frequency, but for interpreters the single most frequently mentioned aspect which could improve interpreter-mediated encounters with children was briefing, for other professionals training (for more details see Amato and Mack 2015).

Comments from interpreters highlight their desperate need to have information about the case and the child before the interview in order to prepare, to know what to expect and to have context to be able to make sense of what is being said. Interpreters should be given a chance to acquire the necessary terminology for the case at hand, but also feel free to accept an assignment if they know they can cope with the case psychologically or turn it down if they feel that it is going to be too stressful for them. Other professionals mentioned training as a priority, which is in line with the complaints about the interpreter’s lack of experience, knowledge and skills they expressed when asked about specific requirements (see 5.3). Other professionals also said, however, that they prefer to work with the same interpreter if possible, and this suggests they feel the need to build a rapport with the interpreter and would be ready to work in a team. This would most probably increase trust and cooperation, something interpreters also demand.

6. Concluding remarks

Rather than drawing our own conclusions, it is more interesting to hear the voice of our respondents about how to meet this (humanitarian) challenge. One comment by an interpreter highlights the need for specific training for both other professionals and interpreters in order to protect the best interest of the child:

(…) moreover police officers should be trained, not only to deal with children, but also to trust interpreters’ skills (…) to find a common approach in order to make the child feel secure. Consequently also interpreters should be trained to act within the boundaries of their task

One of the other professionals wrote:

Training of both other professionals and interpreters: mutual knowledge, knowledge about the child's history and background, agreed interview method and goals as well as an exchange of opinions after the interview

This single statement summarises almost all the sensitive issues when working with children in criminal proceedings: a) need for training on an extensive range of topics for all professionals involved, including interpreters (and we would add: possibly also joint training); b) knowledge, as exhaustive as possible, about the child; c) shared knowledge about interviewing techniques and the goals to be achieved (in other words a briefing with the interpreter); and, last but not least, d) a de-briefing with the interpreter.

We fully agree with the comments quoted above and hope this survey helped to identify some urgent needs in the area of interpreting for children and provided useful indications about actions to improve the situation in Italy in order not only to appropriately train interpreters in the future, but also, and above all, to protect the best interest of the child in line with the UN Convention on the Rights of the Child. For this reason we are very happy that a CO-MINOR II project is already on the way and hopefully will start catering for these needs.


Normative texts

United Nations (1989) Convention on the Rights of the Child (CRC). General Assembly resolution 44/25 of 20 November 1989. URL: http://www.unicef.org/crc/ (accessed 3 December 2016).

European Union (2000) Charter of Fundamental Rights of the European Union (2000/C 364/01). URL: http://www.europarl.europa.eu/charter/pdf/text_en.pdf (accessed 3 December 2016).

European Parliament and European Council (2010) Directive 2010/64 on the right to interpretation and translation in criminal proceedings. URL: http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2010:280:0001:0007:EN:PDF (accessed 3 December 2016).

European Parliament and European Council (2011) Directive 2011/36/EU on prevention and combating trafficking in human beings and protecting its victims. http://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32011L0036&from=IT (accessed 3 December 2016).

European Parliament and European Council (2012) Directive 2012/29/EU establishing minimum standards on the rights, support and protection of victims of crime. URL: http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2012:315:0057:0073:EN:PDF (accessed 3 December 2016).

European Parliament and European Council (2016) Directive 2016/800 on procedural safeguards for children who are suspects or accused persons in criminal proceedings. URL: http://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32016L0800&from=EN (accessed 3 December 2016).

European Commission Recommendation of 27 November 2013 on procedural safeguards for vulnerable persons suspected or accused in criminal proceedings (2013/C 378/02). URL: http://eur-lex.europa.eu/legal-content/IT/TXT/PDF/?uri=CELEX:32013H1224%2802%29&from=en (accessed 3 December 2016).

Official Journal of the European Union 2010/C 115/1, (4.5.2010) The Stockholm Programme: An open and secure system serving and protecting citizens. URL: http://www.eurojust.europa.eu/doclibrary/EU-framework/EUframeworkgeneral/The Stockholm Programme 2010/Stockholm-Programme-2010-EN.pdf (accessed 4 December 2016).

Secondary sources

Alvarado-Little, Wilma (2004) “What did he say? Can you understand that? Interpreting for a child: What training, if any is involved?”. Paper presented at the Critical Link 4 Conference, Stockholm, 20-23 May, 2004.

Amato, Amalia, and Mack, Gabriele (2015) “Briefing, debriefing and support”, in Children and justice: Overcoming language barriers, Katalin Balogh and Heidi Salaets (eds), Cambridge, Antwerp, Portland, Intersentia: 243-312.

Balogh, Katalin, and Salaets, Heidi (eds) (2015) Children and justice: Overcoming language barriers, Cambridge, Antwerp, Portland, Intersentia.

Böser, Ursula, La Rooy, David and Wilson, Christine (2014) “Obtaining interpreter-mediated evidence from children in legal contexts”, in Scottish Institute for Policing Research Annual Report 2013: 34-5. URL: http://www.sipr.ac.uk/downloads/SIPR_Annual_Report_13.pdf (accessed 4 December 2016).

Chand, Ashok (2005) “Do you speak English? Language barriers in child protection social work with minority ethnic families”, The British Journal of Social Work 35, no. 6: 807-21.

Classroom Interpreting. URL: http://www.classroominterpreting.org/ (accessed 4 December 2016).

Dipartimento per la Giustizia Minorile e di Comunità - Servizio statistico (2016) I Servizi della Giustizia minorile. 15 settembre 2016. Dati statistici. URL: https://www.giustizia.it/resources/cms/documents/c7bff0d8e12791ef5c291cb0bec5b602.pdf (accessed 3 December 2016).

European Commission Staff Working Document (2013) Impact assessment accompanying the document ‘Proposal for a directive of the European Parliament and of the Council’ – ‘Proposal for measures on special safeguards for children suspected or accused in criminal proceedings. Brussels, 27.11.2013 SWD(2013) 480 final. URL: http://ec.europa.eu/justice/criminal/files/swd_2013_480_en.pdf (accessed 3 December 2016).

Giovannetti, Monia (2016) VI Rapporto ANCI-CITTALIA 2016 – I comuni e le politiche di accoglienza dei minori stranieri non accompagnati. Un’analisi longitudinale a guida dei percorsi futuri. URL: http://www.camera.it/temiap/allegati/2016/10/19/OCD177-2458.pdf (accessed 3 December 2016).

Granhagen Jungner, J. et al. (2016) “Creating a meeting point of understanding: Interpreters' experiences in Swedish childhood cancer care”, Journal of Pediatric Oncology Nursing 33, no. 2: 137-45 (accessed 4 December 2016).

Green Paper from the Commission (2003) Procedural safeguards for suspects and defendants in criminal proceedings throughout the European Union, Brussels, 19.2.2003 COM(2003) 75 final. URL: http://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52003DC0075&from=EN (accessed 3 December 2016).

Hegemann, Thomas (2009) “Interkulturelle Kompetenz. Basisqualifikation für Arbeit in der Migrationsgesellschaft”, in Kinder zwischen den Kulturen, Reiner Frank (Hrsg.), München, Hans Marseille Verlag, 33-42.

Hiltz, Barbara, and Anderson, Jennifer (2003) “Using interpreters in forensic interviews”, NCPCA Update Newsletter 16, no. 1, American Prosecutors Research.

Jarkman Björn, Gunilla (2005) “Ethics and interpreting in psychotherapy with refugee children and families”, Nordic Journal of Psychiatry 59, no. 6: 516-21.

Jarkman Björn, Gunilla (2013) Refugee children and families: psychological health, brief family intervention and ethical aspects, Dissertation. Linköping University, Department of Clinical and Experimental Medicine, Child and Adolescent Psychiatry.

Justicia European Rights Network (2013) Interim position paper on vulnerable suspected and accused persons in criminal proceedings reform presented by the European Commission on 27 November 2013, URL: http://eujusticia.net/images/uploads/pdf/Interim_JUSTICIA_Paper_on_Vulnerable_Suspected_and_Accused_Persons.doc (accessed 3 December 2016).

Keselman, Olga (2009) Restricting participation: unaccompanied children in interpreter-mediated asylum hearings in Sweden, Linköping Studies in Arts and Science No. 501. Studies from the Swedish Institute for Disability Research No. 31, Linköping, Linköpings universitet.

Keselman, Olga, Cederborg, Ann-Christin, and Linell, Per (2010) “"That is not necessary for you to know!": Negotiation of participation status of unaccompanied children in interpreter-mediated asylum hearings”, Interpreting 12, no. 1: 83-104.

Kirmayer, Laurence J. et al. (2001) Development and evaluation of a cultural consultation service in mental health - Health Transition Fund Final Report, Montreal (Quebec), Culture & Mental Health Research Unit Institute of Community & Family Psychiatry Sir Mortimer B. Davis-Jewish General Hospital, Montreal.

Kriz, Katrin, and Skivenes, Marit (2010) “We have very different positions on some issues: Communication challenges in child welfare work with minority ethnic families in Norway and England”, The British Journal of Social Work 40, no. 5: 1353-67.

Leanza, Yvan, and Rocque, Rhea (2015) “Pediatric settings”, in Routledge encyclopedia of interpreting studies, Franz Pöchhacker (ed.), London/New York, Routledge: 307-9.

Lefevre, Michelle (2010) Communicating with children and young people: Making a difference, London, The Policy Press.

Linell, Per, and Keselman, Olga (2010) “Trustworthiness at stake: Trust and distrust in investigative interviews with Russian adolescent asylum-seekers in Sweden”, in Trust and conflict: Representation, culture and dialogue, Ivana Marková and Alex Gillespie (eds), London, Routledge, 156-80.

London Safeguarding Children Board (2016) London child protection procedures – 6. Working with interpreters / communication facilitators, 5th edition. URL: http://www.londoncp.co.uk/chapters/working_interp.html (accessed 4 December 2016).

Loshak, Rosemary (2003) “The role of the interpreter in child mental health: The changing landscape” in Working with interpreters in mental health, Rachel Tribe and Hitesh Raval (eds), Hove/New York, Brunner-Routledge: 151-67.

Matthias, Carmel, and Zaal, Noel (2002) “Hearing only a faint echo? Interpreters and children in court”, South African Journal on Human Rights 18, no. 3: 350-71.

NET FOR U (2014) Project NET FOR U - NEeds Tackling and NETworks Tracing FOR Unaccompanied minors integration. Integration of third country nationals. HOME/2011/CFP/EIFX/CA/4000001991. Report Italy – Istituto Don Calabria (long version). URL: http://www.oijj.org/sites/default/files/italy_report_long.pdf (accessed 4 December 2016).

Nilsen, Anne Birgitta (2015) “Interpreted communication with children in public-sector services“, Translation and Interpreting 7, no. 3: 121-31.

Norström, Eva, Gustafsson, Kristina, and Fioretos, Ingrid (2011) The interpreter - a cultural intermediary. Behind closed doors. The importance of interpreting for the rule of law and for integration, with special focus on separated minors. Department of Arts and Cultural Studies, Lund University. URL: http://www.tolkprojektet.se/Report_English.pdf (accessed 4 December 2016).

Owens, Robert E. jr (2015) Language development: An introduction, Boston, Allyn and Bacon. First edition 1984, Harlow, Longman Higher Education.

Penn, Claire, and Watermeyer, Jennifer (2014) “Features of cultural brokerage in interpreted child psychiatry interactions: A case of paradoxical practice”, The Interpreter and Translator Trainer 8, no. 3: 354-73.

Phoenix Children's Hospital (2008) Interpreting for children - Spanish bilingual assistant - introduction to medical interpreting. URL: http://dx.confex.com/dx/8/webprogram/Handout/Paper1692/Interpreting%20for%20Children.pdf (accessed 4 December 2016).

Pollitt, Kyra, and Little, Tina (2003) “Beyond the courtroom: Legal interpreting in child protection“, in Traducteurs et interprètes certifiés et judiciaires: Droits, devoirs et besoins, Elena de la Fuente (ed.), Paris, UNESCO/Federation Internationale des Traducteurs: 61-78.

Quindeau, Ilka, and Rauwald, Marianne (Hrsg.) (2016) Soziale Arbeit mit unbegleiteten minderjährigen Flüchtlingen. Traumapädagogische Konzepte für die Praxis, Weinheim, Beltz Juventa.

Raval, Hitesh (1996) “A systemic perspective on working with interpreters“, Clinical Child Psychology and Psychiatry 1, no. 1: 29-43.

Raval, Hitesh (2005) “Being heard and understood in the context of seeking asylum and refuge: communicating with the help of bilingual co-workers“, Clinical Child Psychology and Psychiatry 10, no. 2: 197-217.

Rousseau, Cécile, Measham, Toby, and Moro, Marie-Rose (2011) “Working with interpreters in child mental health“, Child and Adolescent Mental Health 16, no. 11: 55-9.

Rozzi, Elena (2013) Minori stranieri non accompagnati privi di protezione. Ricerca condotta a Torino nell'ambito del progetto PUCAFREU - Promoting unaccompanied children's access to their fundamental rights in the EU. URL: http://www.mshs.univ-poitiers.fr/migrinter/PUCAFREU/ PUCAFREU%20rapporto%20Italia%20IT.pdf (accessed 4 December 2016).

Salaets, Heidi, and Balogh, Katalin (forthcoming 2017) “Participants’ and interpreters’ perception of the interpreter’s role in interpreter-mediated investigative interviews of minors: Belgium and Italy as a case“, in Ideology, ethics and policy development in public service interpreting and translation, Carmen Valero-Garcés and Rebecca Tipton (eds), Bristol, Multilingual Matters: 151-78.

Smith, Melissa B. (2015) “Interpreting in education”, in The Routledge handbook of interpreting, Holly Mikkelson and Renée Jourdenais (eds), London/New York, Routledge: 265-79.

Teke, Gülay (2016) “Migrationssensibler Kinderschutz”, in Bindung und Migration, Karl-Heinz Brisch (ed.), Stuttgart, Klett-Cotta: 101-58.

Together for Short Lives (n.d.) Guidance on working with interpreters in a children’s hospice or a palliative care setting. URL: http://www.togetherforshortlives.org.uk/assets/0000/5959/Working_with_ Interpreters_final.pdf (accessed 4 December 2016).

Van Schoor, Dominique (2013) Interpreter-mediated interviews of child witnesses and victims: Status quaestionis, Master thesis - Faculteit Letteren Taal en Communicatie Antwerpen. URL: http://www.arts.kuleuven.be/hub/tolkwetenschap/projecten/co_minor_in_quest/interpreter-mediated-interviews-of-child-witnesses-and-victims-status-quaestionis (accessed 4 December 2016).

Viràg, György (2015) “Interpreted interviews with highly vulnerable children”, in Children and justice: Overcoming language barriers, Katalin Balogh and Heidi Salaets (eds), Cambridge, Antwerp, Portland, Intersentia: 77-93.

Winter, Karen (2010) Building relationships and communicating with young children. A practical guide for social workers, London, Routledge.


[1] Sections 1, 2, 4.2, 4.3, and 5.2 to 5.4 by A. Amato; sections 2.1 to 4.1, 5, 5.1, and 6 by G. Mack.

[2] In practically all international law texts a child is a person under 18 years of age.

[3] The Stockholm Programme is a five-year plan (2010-2014) on justice and home affairs in the EU member states. It “defines strategic guidelines for legislative and operational planning within the area of freedom, security and justice in accordance with Article 68 TFEU.“ (Official Journal of the European Union 2010/C 115/1, 4.5.2010, Art. 1., last paragraph).

[4] For a more detailed discussion on children and vulnerability in legal contexts see Viràg (2015).

[5] For a more detailed description of this part of the research see chapter 4 in Balogh and Salaets (2015: 175).

[6] All translations of comments by respondents quoted in the following are ours.

About the author(s)

Amalia Amato

Gabriele Mack is associate professor of German language and translation at the Department of Interpreting and Translation of Bologna University at Forlì, where she teaches interpreting between German and Italian. She has been working as a free-lance conference interpreter and translator for many years. Her main research interests focus on interpreter training, analysis of interpreting processes and outcomes as well as conference, media and public service interpreting.

Email: [please login or register to view author's email address]

©inTRAlinea & Amalia Amato & Gabriele Mack (2017).
"Interpreters Working with Children in Italy Profile, Role and Expectations"
inTRAlinea Volumes
Edited by: {specials_editors_volumes}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2263

The Use of Respeaking for the Transcription of Non-Fictional Genres

An Exploratory Study

By A. Matamala, P. Romero-Fresco & L. Daniluk (Universitat Autònoma de Barcelona, Universidade de Vigo, Spain & University of Roehampton, UK)

Abstract & Keywords

Transcription is not only a useful tool for audiovisual translation, but also a task that is being increasingly performed by translators in different scenarios. This article presents the results of an experiment in which three transcription methods are compared: manual transcription, respeaking, and revision of a transcript generated by speech recognition. The emphasis is put on respeaking, which is expected to be a useful method to speed up the process of transcription and have a positive impact on the transcribers’ experience. Both objective and subjective measures were obtained: on the one hand, the time spent on each task and the output quality based on the NER metrics and on the other, the participants’ opinions before and after the task, namely the self-reported effort, boredom, confidence in the accuracy of the transcript and overall quality.

Keywords: audiovisual translation, subtitling, respeaking, speech recognition, transcription

©inTRAlinea & A. Matamala, P. Romero-Fresco & L. Daniluk (2017).
"The Use of Respeaking for the Transcription of Non-Fictional Genres An Exploratory Study"
inTRAlinea Volumes
Edited by: {specials_editors_volumes}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2262

1. Introduction

A written transcript of an audiovisual product may be needed in various scenarios: for instance, audiovisual translators (whether for subtitling, dubbing or voice-over) often require a written script of the audiovisual content they are translating in order to reduce costs and increase translation accuracy and speed. Whereas up until now this job was only performed by professional transcribers, translators are increasingly being asked to produce transcriptions (Cole 2015). As for the notion of accessible filmmaking (Romero-Fresco 2013), which envisages the inclusion of translation and accessibility from production, companies such as Silver River, Endemol or Subtrain in the UK are beginning to employ translators to first transcribe and then translate documentaries. Furthermore, translators are also being hired to provide transcriptions for EU-funded projects on speech recognition-based subtitling such as SAVAS (http://cordis.europa.eu/fp7/ict/language-technologies/project-savas_en.html, see below), and transcriptions obtained using different methods (automatic/semi-automatic/manual, professional/non-professional) are also generated to process large speech corpora in research environments.

Professional transcribers often provide high-quality manual transcripts. However, this paper aims to explore whether alternative processes based on speech recognition systems, either automatic or via respeaking, could be successfully used when generating a written transcript of an audiovisual product in a professional environment. The focus will be on non-fictional content that has to be voiced-over in the target language and has not been previously scripted. Our hypothesis is that respeaking could be successfully used to speed up the process of transcription and could have a positive impact on the transcribers’ experience.

This exploratory research is linked to the three-year project “ALST-Linguistic and Sensorial Accessibility: Technologies for Voice-over And Audio Description”, funded by the Spanish Ministry of Economy and Competitiveness (reference code FFI-2012-31024) in which the implementation of three technologies (speech recognition, machine translation and speech synthesis) in two audiovisual transfer modes (audio description and voice-over) was investigated. The focus in this exploratory study is on speech recognition as applied to the transcription of voice-over, with special emphasis on respeaking.

The article begins with brief definitions of some of the key concepts used in our research, especially of what transcribing means (section 2), and then offers a short overview of previous research on respeaking, mostly as a transcription tool (section 3). It must be stressed that, even though three transcription techniques are tested in the experiment (manual transcription, respeaking, and revision of an automatically-generated transcript), our interest lies mainly in respeaking, since our hypothesis is that this technique will yield better results. The third section is therefore devoted to this technique and does not address transcription via other procedures. Finally, the article describes the experimental set-up of the current test and discusses its results.

2. Transcribing: different needs, different approaches

Transcription is often regarded simply (and simplistically) as the written representation of audible speech. This definition fails to account for the complexity involved in a phenomenon that, as demonstrated by Ochs (1979) in her seminal work, is theoretical in nature, “a selective process reflecting theoretical goals and definitions” (ibid. 44). Indeed, despite the inadequate attention given to transcription in qualitative research (Oliver et al. 2005), there is now agreement that transcribing is an interpretive act rather than simply a technical procedure (Bailey 2008). According to Davidson (2009), transcription may thus be defined as follows:

a representational process (Bucholtz, 2000; Green et al.,1997) that encompasses what is represented in the transcript (e.g., talk, time, nonverbal actions, speaker/hearer relationships, physical orientation, multiple languages, translations); who is representing whom, in what ways, for what purpose, and with what outcome; and how analysts position themselves and their participants in their representations of form, content, and action. (Green et al., 1997, p. 173)

An example of the complexity of transcription is found in projects creating corpora from speech data, as put forward by Bendazzoli and Sandrelli (2007), who state that transcribing was the most challenging task in their study. Some of the aspects that need to be clearly defined when preparing transcription protocols are the level of detail in the annotation (linguistic, paralinguistic, non-verbal and visual information, for instance), the spelling and punctuation conventions to be applied, and the technology to be used (Bailey 2008). Indeed, transcribing implies making choices and this selectivity of the transcripts is always related to the goals of the study, as “it is impossible to record all features of talk and interaction” (Davidson 2009: 38).

Taking into account the aim of our research, transcription is not viewed as a process to generate annotated data for research purposes but as a professional task in which a written text needs to be created from a non-scripted audiovisual source. Hence, participants were expected to follow the standard procedures in the video production industry, which includes a written representation of audible speech, the identification of the person speaking and relevant paralinguistic features such as <laugh> that can facilitate the editing process. No further annotations were requested, as these were the ones a professional audiovisual translator working on a non-fictional content to be voiced-over would need.

Voice-over was indeed the audiovisual transfer mode chosen for the experiment, and can be defined as follows:

Technique in which a voice offering a translation in a given target language is heard simultaneously on top of the SL voice. As far as the soundtrack of the original program is concerned, the volume is reduced to a low level that can still be heard in the background when the translation is being read. It is common practice to allow the viewer to hear the original speech in the foreign language at the onset of the speech and to reduce subsequently the volume of the original so that the translated speech can be inserted. The translation usually finishes several seconds before the foreign language speech does, the sound of the original is raised again to a normal volume and the viewer can hear once more the original speech. (Díaz-Cintas and Orero 2006:473)

Voice-over is used in many countries to revoice non-fictional content, but it is also used to revoice fictional genres in Eastern European countries such as Poland. Although many voice-over translators work without a script, the availability of an accurate transcript reduces the time and increases the accuracy of the translations (Franco et al. 2010).

As said before, transcription of non-fictional content for voice-over has been generally performed manually but two alternative systems are investigated in this research: respeaking and revision of automatic speech recognition transcript. A brief definition of both is offered next.

Respeaking is defined by Romero-Fresco as a technique in which a respeaker listens to the original sound of a live programme or event and respeaks it, including punctuation marks and some specific features for the deaf and hard-of-hearing audience, to speech recognition software, which turns the recognized utterances into subtitles displayed on the screen with the shortest possible delay. Respeaking has become one of the preferred methods for creating live subtitles, although other methods have been used in the past and are sometimes still used: standard QWERTY keyboards, Velotype, tandem or rotation keyboards, and stenotype (Romero-Fresco 2011:1-13). However, in this article we consider respeaking a technique in which the original verbal aural elements of an audiovisual product (in our experiment, recorded interviews) are repeated or rephrased by a professional transcriber to speech recognition software in order to generate a live transcript of the original dialogues/monologues (in our experiment, with a view to creating a translation using voice-over). This is similar to what in the USA is called (real-time) voice-writing/reporting and is used to produce verbatim reports in court, an area in which speech technologies have also been researched (Sohn 2004). This technique has also been used in the EPIC project to create corpus transcripts (Bendazzoli and Sandrelli 2005).

As for automatic speech recognition (ASR), also known as computer speech recognition or simply speech recognition, Anasuya and Katti (2009:181) present a review of the past sixty years of research on the topic and define it as “the process of converting a speech signal to a sequence of words, by means of an algorithm implemented as a computer program”. In this case, the process is fully automatic with no dictation, but human revision is expected when certain quality thresholds are to be met. This is, in fact, the approach taken in our study.

2. Previous research on respeaking

Respeaking as a tool for the production of live subtitles has been the subject of increasing research in recent years, with Eugeni (2008a; 2009) and Romero-Fresco (2011) being two of the main authors in the field. Going back a few years, an excellent forum to gain an overview of incipient research on respeaking from the perspective of audiovisual translation was the International Seminar on New Technologies in Real Time Intralingual Subtitling, held in Forlì in 2006.[1] Most of the contributions focused on the process and professional practice of respeaking as implemented in different countries. The second conference,[2] held three years later in Barcelona focused on the reception of respoken subtitles and the strategies used by live subtitlers. The next conference in the series, held in Antwerp in 2011,[3] approached respeaking from the point of view of system developers, broadcasters and academics. The 2013 conference in Barcelona[4] featured the presentation of the Translectures project and the EU Bridge project, both of which looked into the automatic transcription and translation of online videos, lectures and webinars. Finally, the latest conference, held in Rome in 2015,[5] explored issues such as live subtitling quality on TV, respeaking by blind professionals and experimental studies on the reception of respeaking.

Beyond the scope of this conference series and others,[6] the need to train respeakers has generated several didactic proposals. These contributions often link the competences of respeakers to those of simultaneous interpreters (Eugeni 2008b; Arumí-Ribas and Romero-Fresco 2008; Romero-Fresco 2012a), given the multitasking and split-attention requirements of both disciplines.

The process of respeaking has also attracted the attention of researchers such as Van Waes, Leijten and Remael (2013), who have investigated the reduction strategies used by live subtitlers. Research on viewers’ comprehension of respoken subtitles has also been conducted. Romero-Fresco (2010; 2012b) reports on the results of two experiments: the first aimed to test the comprehension of respoken subtitles for the news by hearing, hard-of-hearing and deaf viewers. Subtitles were presented at two different speeds and results show that most viewers find it difficult to obtain enough (verbal and visual) information from respoken subtitles with no sound. The second experiment was carried out using an eye-tracker and aimed to research how viewers read respoken word-for-word subtitles as opposed to block subtitles. The results were significantly better for the latter than for the former.

More recently, research has been carried out on the quality of live subtitles, looking at the potential use of metrics. Traditional Word Error Rate (WER) methods have been adapted to the specificities of respeaking with models such as that developed by CRIM in Canada (Dumouchel et al. 2011) and the NER model (Romero-Fresco and Martínez 2015), which is being used by researchers, governmental regulators and broadcasters in different countries around the world (Ofcom 2013; MAA 2014). In 2013, the British regulator Ofcom commissioned a two-year research project to assess the quality of live subtitles on UK TV. This has resulted in the largest analysis of live subtitling available to date, providing data on the accuracy, speed, delay and edition rate of the live subtitles produced for 50 hours of live TV material, including 507,000 words and 78,000 subtitles (Romero-Fresco 2016).

Another leading project on respeaking is the EU-funded SAVAS project (2012-2014) (http://cordis.europa.eu/fp7/ict/language-technologies/project-savas_en.html), which aimed to develop speech recognition technology for live subtitling in different languages and in the domain of broadcast news, sports, interviews and debates. The project dealt with seven languages and created solutions for a speaker-independent transcription system, a collaborative respeaking-based system for live subtitling and the batch production of multilingual subtitles.  

However, despite the interest raised by respeaking in the industry and amongst researchers, there is not much literature regarding the implementation of respeaking in areas other than TV, such as live events, or the comparison of its efficiency against ASR or manual transcription, which is currently the most common method used to transcribe non-fictional content. Sperber et al. (2013) propose a method for efficient off-line transcription of speech through respeaking via a combination of various techniques. Their method segments the speech input into short utterances, and selects only some of the utterances for respeaking based on confidence measure estimates. They present results from experiments by two respeakers and a simulation, and demonstrate the potential of their method in increasing speed and correction efficiency. However, the selection of respeaking, or typing as the better choice, is highly dependent on the particular segments. Bettinson (2013), on the other hand, investigates the application of respeaking to the transcription of data by field linguists, although for him transcribing means producing oral annotations (namely “careful speech” of the same content), a phrase level translation into a language of wider communication plus analytical comments. Also, for research purposes, Bendazzoli and Sandrelli (2007) use speech recognition software programs such as DNS or IBM Via Voice to obtain draft data which are later revised. The process explained in their study, which is referred to in the paper as shadowing because they adopt an interpreters’ perspective, consists in listening to the recording and repeating aloud what the speaker says.

In the medical field, and mostly in the USA, researchers have looked into the feasibility of replacing manual transcription of patients’ data with a semi-automatic approach based on respeaking (or voice writing, as this technique is known there). Al-Aynati et al. (2003) found transcriptions generated by dictation with speech recognition software to be a viable tool to transcribe pathology reports but also more time consuming than manual transcription, given the extra time needed to edit the errors caused by the speech recognition software. More recently, and taking advantage of more accurate software, Singh and Pal (2011) compared the feasibility and impact of making a transition from a manual transcriptionist-based service to one based on dictation with speech recognition in surgical pathology. The results of their experiment showed significant improvements in turnaround time and a positive impact on competency-based resident education. These findings have been echoed and confirmed in more recent studies (Joshi et al. 2014, Hartman 2015) which stress the usefulness of speech recognition-based transcription in the medical field.

Perhaps the most important study so far is the small-scale test conducted by D'Arcangelo and Cellini (2013) comparing the speed in the transcription of three samples of fast, medium-paced and slow speech using four methods: manual transcription, stenography, respeaking and automatic recognition plus correction. Although the test only offers anecdotal evidence, given that it compares the performance of two participants, it provides interesting data. Automatic recognition plus revision performs well in the transcription of slow speech and significantly worse when dealing with fast speech. As for the other methods, manual transcription is normally slower than respeaking and stenotyping, which is the fastest method.

Taking into account these encouraging results and the improvement in accuracy experienced by speech recognition software over the past years, the present article attempts to test speech recognition-based transcriptions in the film industry, which has so far relied extensively (and, to our knowledge, almost exclusively) on manual transcription (Pollak 2008).

3. Experimental set-up

The aim of the experiment presented in this article was to gather quantitative and qualitative data to compare three scenarios in the transcription of a non-fictional audiovisual excerpt: manual transcription, respeaking, and the revision (or post-editing) of a transcript generated by ASR. As stated in the introduction, our main interest was in the implementation of respeaking, hence our hypothesis was linked to this technique. A pilot test was carried out with five participants to validate the experimental set-up described in this section.

3.1. Participants

Ten native English participants (4 male, 6 female), all professional transcribers, took part in the experiment, which obtained ethical approval from the Universitat Autònoma de Barcelona. The quantitative data from two participants and the qualitative data from one participant could not be used due to technical reasons. All participants had professional experience in manual (QWERTY) transcription and none of them had used respeaking to transcribe audiovisual content before. Although an experiment with different profiles (for instance, professional respeakers versus professional transcribers) would have yielded interesting results, it was beyond the scope of this project and a homogeneous sample of participants was prioritized. In this regard, all of them were working as transcribers for independent companies that produce documentaries, reality TV shows, factual entertainment, as well as comedy and entertainment programmes. In terms of tools, they reported having used dedicated software such as Forscene, Aframe, Avid Interplay or Synergy but very often they simply used software to preview video files like VLC player and a word processing tool. Only one participant had used speech recognition software (DNS) to complete university assignments, but not for transcription purposes. The participants indicated different levels of satisfaction with their current jobs, 3.14 being the average on a 5-point Likert scale.

3.2. Materials

A 12-minute video interview was split into three four-minute excerpts, equivalent in terms of number of words. The video featured two female American Hip-Hop artists from California, Gavlyn and O’Blimey, as well as Kasia Ganzera, a young journalist working for an online portal called ‘Hip-Hop Says’. The content of the interview was related to the recent work of the two artists, their European tour, their past experiences, and their future plans and aspirations. The language spoken was colloquial English, including many slang words (see annex for the transcriptions). This material was chosen by one of the authors of this study, himself a professional transcriber, as a representative example of material for transcription, featuring a real-life situation with unstructured, spontaneous colloquial speech for which no script is usually available.

Participants used Dragon Naturally Speaking 12 Premium to produce the respoken transcription of the video. In order to test the ASR method (automatic transcription plus manual revision), a comparative analysis was performed between the automatic transcription tool in Dragon and the state-of-the-art EML transcription server. The latter produced better results and was therefore chosen to generate the transcript for this part of the experiment. Even though the fact that this ASR system had not been trained specifically for the audiovisual content at hand may have an impact on the accuracy of the automatic transcription, it is an important requirement to test how a non-trained engine would perform with spontaneous speech.

Two pre-test questionnaires were prepared: one to gather demographic information at the beginning of the session and another one to find out about the participants’ experience and views prior to the tasks. A post-test questionnaire was also prepared including questions on self-reported effort, boredom, confidence in the accuracy of the transcript and overall quality of each of the tasks (on a 5-point Likert scale), as well as subjective questions about the three transcription techniques considered here.

3.3. Procedure

The participants were welcomed in a computer lab and were informed about the general structure of the experiment. They filled in the pre-test demographic questionnaire and were provided with a 30-minute training session on respeaking by a professional respeaking trainer. After the session, they filled in the second pre-test questionnaire and proceeded to watch the video content. Each participant was then asked to transcribe three excerpts from the video interview using three methods: manual transcription, respeaking and revision of transcript generated by ASR. The order of the tasks and the videos used for each task was randomized and balanced across participants. In order to avoid fatigue, a maximum duration of 30 minutes was set per task, with a total of 90 minutes for the whole experiment. Those participants who managed to complete their transcription in 30 minutes or less were asked to note down the exact time spent on each task, whereas those who did not manage to complete their task noted down the exact time-code where they had stopped after 30 minutes. After completing the three tasks, the participants filled in the post-test questionnaire. Two researchers were in the room to monitor the participants and take observational notes. This was useful to discover, for instance, the mixed approach adopted by some of the participants who started to correct the ASR transcript, as discussed in section 4.

3.4. Methodology

Quantitative data was obtained on the time spent on each task as well as the ratio of minutes spent on the transcription per minute of original content. Qualitative data was also obtained via the questionnaires and the observation and interaction with participants. Descriptive statistics were used for the analysis.

The transcripts generated by all the participants were also assessed against a reference standard using an adapted version of the NER model (Romero-Fresco and Martínez 2015) and the NERStar tool (http://www.speedchill.com/nerstar/). The aim of this analysis was to detect the influence, if any, of the transcription method on the quality of the final output. The NER model defines accuracy with the following formula:


N refers to the number of words in the transcription (or subtitles), including commands and words. E refers to edition errors, generally caused by the subtitlers’/transcribers’ decisions. R stands for recognition errors, meaning misrecognitions caused by mispronunciations or mishearings. E and R errors can be classified as serious, normal or minor, scoring 0.25, 0.5 and 1 respectively. For the purpose of this study and given that the NER model is designed for the analysis of live subtitles, the model was stripped of some of its subtitling-specific components, such as those that account for issues of delay, positioning and on-screen display. The different degrees of severity were still found to be valid and useful for the analysis of transcription as was the distinction between E and R. The latter accounted in this case for recognition and typing errors in the case of manual transcription.

4. Results and discussion

Regarding the time spent on the experiment, only three participants out of eight managed to complete the task in the required time (a maximum of 30 minutes) with manual transcription and with the revision of the ASR transcript, whilst in respeaking the number increased to five participants out of eight.

Considering only the data of those participants who finished (three from manual and ASR revision, five from respeaking), manual transcription is the fastest technique and the revision of the ASR transcript is the slowest one (see Table 1). As for the quality of the output as calculated with the NER metrics, the highest quality was obtained with the revision of the ASR transcript, followed by manual transcription and respeaking. It was observed that some participants started off correcting the automatically-generated transcript but then deleted the remaining content and transcribed it manually. For this reason, figures about the revision of the ASR script are often closer to those of manual transcription. Future research with higher quality ASR (or with less spontaneous material) may yield different results, but this behavior may provide interesting insight into the transcribers’ attitude towards ASR depending on its accuracy rate, all of which could be analysed with keylogging software. Respeaking has the lowest NER values, which may be due to the fact that most participants did not perform a thorough revision of the respoken text. Instead, they just provided the raw respoken results. In any case, the NER score in all the conditions is relatively low if compared, for instance, with the minimum score for subtitles to be considered as acceptable, which is 98% (Ofcom 2015). Yet, as mentioned in the introduction, while live subtitles are produced by fully qualified respeakers, using a trained speech recognition software to generate an output that is used by billions of viewers, the objective of transcription is usually different, as different are the conditions in which this experiment was carried out. It is for this reason that 98% may be seen as an excessively high threshold for transcription, and further research on metrics for this type of transcription technique should be carried out.





Speed (time spent on one minute of programme)




Accuracy (NER)




Table 1: Mean data from participants who finished the task

When considering all the participants (see Table 2), manual transcription was again the fastest option, followed by respeaking and the revision of the ASR transcript. Once again, some participants felt frustrated with the revision of the automatic output, deleted all the content and manually transcribed the excerpt, which is a direct consequence of our experimental design and should be taken into account in future research.

Concerning NER values, manual transcription yields the highest scores, followed by the revision of the ASR transcript and respeaking. None of them, however, reaches the above-mentioned 98% threshold.





Speed (time spent on one minute of programme)




Accuracy (NER )




Table 2: Mean data from all participants

These data are in line with previous results and show that, overall, manual transcription is the fastest option. This is a logical consequence of the participants’ experience as professional manual transcribers who have developed their own techniques. Accuracy in manual transcription is the highest when considering all the participants and the second highest when considering only those who completed the task. Once again, their professional practice with manual transcription generally involves revision. What is, however, surprising, is that in manual transcriptions their accuracy is below 98%, which in other fields such as subtitling has been established as the minimum NER score to provide a quality output. This shows again that research is needed on what the minimum score for a quality transcription should be.

Respeaking was the second fastest in the experiment. It is the method that allowed the highest number of participants to finish, but it also slowed down other participants, thus showing a wide range of speeds. It obtained the lowest results in terms of accuracy, mainly because most participants did not revise the respoken transcript. These results show the potential of this technique for the transcription of audiovisual material but also two aspects that must be factored in the equation: training and time for revision. The participants had a 30-minute crash course on the basics of respeaking, while the average training for respeakers ranges from 3 weeks to 3 months (Romero-Fresco 2012a). Future research will also need to account for the impact of revision on the overall speed of transcription, which may be higher with respeaking and ASR than with manual transcription.

As for the revision of an ASR transcript, the results are very much determined by the approach adopted by the different participants. Most of them began by attempting to correct the ASR transcript, but then decided to delete it and change to manual transcription. It then became a type of mixed approach, a slower version of manual transcription with higher accuracy because revision was at the heart of it. However, the increase in time did not result in an increase in quality in all cases. Although quality was higher than with manual transcription and respeaking when considering only the participants who completed the task in 30 minutes, it was lower than with manual transcription when considering all the participants

Regarding qualitative and more subjective data, Table 3 includes a comparison between the participants’ opinion on ASR and respeaking before the experiment (with no knowledge or experience on ASR and respeaking other than the 30-minute crash course they had just received) and after the experiment. Table 3 includes the statements with which they had to indicate their level of agreement on a 5-point Likert scale, next to the pre-task and post-task mean data.


Pre-task mean

Post-task mean

Manual transcribing is too time consuming



Respeaking could be a useful tool to transcribe documentaries



Automatic speech recognition could be a useful tool to transcribe documentaries.



Respeaking could speed up the process of transcription



Automatic speech recognition could speed up the process of transcription



Respeaking could increase the accuracy of transcriptions



Automatic speech recognition could increase the accuracy of transcriptions



Respeaking could increase the overall quality of transcriptions



Automatic speech recognition could increase the overall quality of transcriptions.



Table 3: Pre-task and post-task opinions (mean data on a 5-point Likert scale)

Before the experiment, participants expected respeaking to be a very useful technique to transcribe documentaries (4.5) and to increase the speed (4.5), accuracy (3.8) and overall quality (3.4) of the transcriptions. These expectations dropped to different extents after the experiment, with 3.8 for usefulness, 3.9 for speed, 2.9 for accuracy and 3.1 for overall quality. In other words, having completed the experiment with a mere 30-minute training session on respeaking, most participants regard this method as useful, fast (especially given that they consider manual transcription, which they do for a living, too time-consuming) and with an acceptable quality. They do have doubts, however, regarding the accuracy of the transcriptions.

In contrast, the expectations raised by ASR were not met after the experiment. On paper, ASR looked useful (4.1) and it was expected to speed up the process (4.1), the accuracy (3.00) and, to a lesser extent, the overall quality (2.8) of the transcriptions. After the experiment, ASR was no longer regarded as useful (2.7), fast (2.1), accurate (2.2) and of good quality (2.5), with all items under analysis below 2.5 on the scale. Interestingly enough, though, the NER metrics show that quality of the ASR transcription did increase, which indicates that it may be the self-reported effort (see Table 4) that makes the participants think that the accuracy of ASR is so low.

After the experiment, participants were also asked about their perceptions in terms of effort, boredom, accuracy and overall quality with regard to the transcripts they generated. Results are summarised in Table 4.





Self-reported effort












Overall quality




Table 4: Participants’ assessment

Respeaking is regarded as the least taxing (2.89) and least boring (2.22) method, in contrast particularly with the revision of an ASR transcript (4.55 and 3.89, respectively) but also with manual transcription (3.11 and 3.12, respectively). In other words, respeaking seems like a very attractive method with which the participants would like to replace manual transcription. However, they still harbour doubts regarding its overall quality (3.22) and especially its accuracy (2.78), which is reasonable given the objective results shown in tables 1 and 2. In these two aspects, manual transcription is still regarded as more reliable (4.22 and 4.33, respectively), even though objective metrics show that in reality the revision of the ASR transcript was more accurate than the manual one.

A final open question asking participants about their general views on the experiment and the possible use of respeaking for the transcription of audiovisual material provided some interesting additional insights, which we summarise next.

Participant 1 was “impressed with the re-speaking software… especially when used in combination with manual input. The video material was using slang language and I feel that re-speaking could work even better in an interview situation with (standard) English and no slang”. As indicated before, this excerpt was chosen because it was particularly challenging and because it is a stereotypical example of extemporaneous audiovisual content. However, as suggested by this participant, other types of content could probably yield even better results. This statement also points indirectly to an interesting area of research, that of automatically assessing what content is worth respeaking and what content should be manually transcribed. In addition, the observation of the participants in the experiment showed what may be the most effective transcription method: a combination of manual transcription and respeaking. Instead of opting for hands-free respeaking, some participants chose to combine respeaking with the occasional use of typing, mostly to correct errors, thus making the most of the speed inherent to respeaking and of their proficient typing skills. Striking the right balance between manual transcription and respeaking could be a good solution to increase the speed and (crucially) the accuracy of transcriptions. This can only be achieved after proper training and practice, none of which could be provided in the experimental setting. Some participants suggested in informal post-test interviews a combination of 75% voice and 25% manual transcription as an ideal balance, but this remains to be verified by further research.

Participant 4 thought respeaking could be useful to improve accuracy for those transcribers who cannot spell very well, whilst stressing that more specific training was required. Participant 5 also highlighted the need for more practice, as did Participant 6, who added some interesting insight concerning his/her job as a manual transcriber: “I am a fast typist but I am used to a different keyboard, so this slowed me down on the manual transcription. I am also used to using a transcription software such as Express Scribe, which makes manual transcription a bit faster. I was a lot slower with Dragon because it takes time to learn how to use the different options such as go back, vocab, etc. I think in the long term this would be much quicker than manual transcription.”

On the other hand, most views on ASR are negative, such as those expressed by Participant 2, who explains that s/he had to re-do everything by hand, or Participant 3, who considers that the technology still needs to improve. It is important to point out that the ASR system was not adapted to the domain or the speakers in the video, which undoubtedly affected the quality of the results. A better performance could be expected when using an engine trained on the specific domain of the audiovisual material tested.

To conclude, the participants were asked to answer the question “Would you enjoy your job more than you enjoy it now if you used respeaking for transcription?” Five possible replies were given on a 5-point Likert scale, and results show that 22.22% would enjoy it much more and 66.66% would enjoy it a bit more, while only 11.11% consider they would have the same level of enjoyment and none considers they would enjoy it less than manual transcription.

5. Conclusions

This article has presented a first exploratory study on the potential use of respeaking as a tool for transcribing non-fictional audiovisual material through an experiment in which three conditions were compared, namely manual transcription, respeaking and the revision of transcript generated by ASR. Although limited in its number of participants, our research has provided both quantitative and qualitative data that will hopefully be useful for, and expanded in, future research.

Manual transcription was the fastest option, followed by respeaking and ASR, but it was respeaking that allowed the highest number of participants to finish the transcription in the time allocated for the task. The downside is that accuracy levels in all conditions were unexpectedly low (and lower than the 98% accuracy rate required in other fields such as subtitling), especially in respeaking. On the one hand, this makes us think that further research is needed to ascertain what the minimum NER score for a quality transcription should be, a value that may be linked to the purposes of such transcription. On the other, this shows that manual transcription is still the most effective method in terms of speed and accuracy, which comes as no surprise taking into account that participants were professional manual transcribers. Still, as shown by subjective opinions gathered via questionnaires, transcribers are fairly tired of the manual method. They regard it as too time-consuming, taxing and boring, and are willing to try out new possibilities. Therefore, the results of the experiment only partially fulfill the initial hypothesis, which suggested that respeaking could be successfully used to speed up the process of transcription and could have a positive impact on the transcribers’ experience.

When presented with alternative methods, they initially embrace the idea of using ASR, but after the test they find it is less useful than expected. In our experiment, the revision of a script generated by ASR does not reduce transcription time because many participants end up doing a manual transcription from scratch after deleting the ASR content. However, further research should be conducted with this method, since the conditions used in this experiment (an ASR engine that was not adapted to the genre of the audiovisual material tested) did not allow the system to reach its standard performance rates.

Respeaking is very welcome by participants and performs very well in terms of speed, even though participants were not familiar with this technique and the training was extremely short and not suited to the specificities of this type of transcription. As indicated by one of the participants after the experiment, respeaking was perceived as a very fast system which could solve spelling issues but, in order to make it accurate, specific training and a clear method would be needed. The training provided was short and focused on producing subtitles, whilst for professional transcribers the specificities of the task at hand should be taken into account. As observed in informal talks after the experiment, it seems that a combination of respeaking and manual transcription could be the solution.

From a research perspective, it would be worth investigating a method with a quality control system to automatically propose to the end user the most efficient transcription method for a given video, including ASR if it proves to be reliable enough. It would also make sense to conduct a more specific study comparing the speed and accuracy of manual transcription versus respeaking taking into account not only the transcription time but also the revision time. Another aspect worth investigating is the effect of longer working sessions, since the results for this experiment were obtained in tasks limited to a maximum of thirty minutes. It remains to be seen how professionals would perceive respeaking in terms of effort and boredom after longer sessions. Experiments with a higher number of participants and different audiovisual material are definitely needed to yield more conclusive results. For instance, in voiced-over non-fictional genres translators usually edit out many features such as repetitions, false starts, hesitations and non-relevant discourse markers. Working with an edited text generated by respeaking could maybe speed up the process or even facilitate machine translation plus post-editing, an aspect worth researching.

Many research possibilities emerge and further experimental studies that can have a clear impact on the industry will hopefully be derived from this first experiment. For the time being, transcribers seem ready to test respeaking in order to make their jobs not only more efficient but also more enjoyable.


This article is part of the project ALST (reference code FFI-2012-31024), funded by the Spanish Ministerio de Economía y Competitividad. Anna Matamala and Pablo Romero-Fresco are members of TransMedia Catalonia, a research group funded by the Catalan Government (reference code 2014SGR0027). We would like to thank all the professional transcribers who took part in the experiment. Special thanks to Juan Martínez and the European Media Laboratory (EML GmbH) for their support.

Appendix: Transcription Excerpts

Clip 1

00:01 L: okay I'm rolling

K: yeah, okay, let's go, hey, what's up guys today we are with Gavlyn and O'Blimey

O: it's good

G: hey


K: guys, so, crazy thing that I saw you 2 years ago in the cipher effect, right, that was like the first one and then I'm like, okay, that's cool, dope stuff, then you're featured on O'Blimey's ill video and on the whole track, and now you are on your European tour

O, G: <laughs>

K: guys, that's crazy, how, how like, how did that happen, like, how comes it took you like two years to kind of do something together?

G: well she's lived in San Francisco for a minute, she just barely moved to Los Angeles and the first time we met was at the cipher effects, so

K: oh, was it? Okay, alright

O: I just moved down to LA


O: and was like living in LA half time, and San Francisco half time, working on my music, so, so really like making time to collaborate with a bunch of people in LA, just was a happening, you know at the time and as soon as I was there full-time, you know I let Gavlyn know that I was back in town and ready to work as well as the other girls from the cipher too, we've all collaborated a bunch, but Gav, I was like I have a song for you right now, she called me on her way back from another Europe tour

K: okay


O: and she was like, you want to get together? You want to work on something? I was like, I have a song and studio time ready to go right now, so she came into the studio that night, right after getting back from the airport and we recorded 'ill' and literally since that song we have hang out like every other day, ever since that point, so

G: yeah

K: okay, so it was cool for you to actually spend so much time together, because I mean European tour, you've got like, I don't know 30 shows or something like that

G: we've got 20

O: 25

G: 25

O: yeah

K: that's crazy long time, but, like how do you feel about this whole tour, because it must be a lot of work and travelling and you guys kind of working together as well, like how does that make you feel


G: it feels awesome, like, I consider O'Blimey one of my best friends so it's super tight to be touring with my girl and kill it and see the people enjoy what we do together, it's an amazing feeling, I love it

O: like, no doubt that there's tonnes of work, but it's kind of like getting to go on holiday or going on vacation, you know

G: yeah

O: you know with your best buds, we may annoy the shit out of Mike Steez our DJ and manager from time to time, but I think the three of us together, we have a good team, and

K: it's good

O2: 21

O: it's just, we vibe, we mesh, you know, it's easy

K: yeah, so, what would you say is the best thing about hip-hop culture, like what's the thing that you love the most about it?

G: you start

O: the best thing about hip-hop culture? Man, I mean, a very, you know, a very wise person once said is that you know the best thing about music, and it goes for hip-hop for me, is that when it hits you you feel no pain

K: yeah


O: and you know, whether or not a track, or a vibe in an area brings you to a painful place, the fact that everybody else, you know, kind of understands and has been there as well, whether or not the context matches perfectly, but just the fact that other people understand what you are going through and there is an entire community that goes through it as well

K: that is a great thought

G: exactly

O: especially like, with the honesty in hip-hop right now, everybody is just telling it like it is and I think that, it's rare, it's beautiful


G: I have to second on what she is saying, like, that it's the coolest thing about hip-hop, it's like just being able to write shit and do shit that, you feeling explains who you are, you know what I mean? And people are enjoying that and vibing out with it, it's super tight

K: cool, okay, I saw your hip-hop Kemp the other time and I was like, okay, crazy show you and Organised Threat on the show and Yella Brother was with you as well, I was like, okay, you literally just just burnt the stage down, then I'm on my way to RA the Rugged Man, I was like yeah, waiting for the show


K: and then I see you guys on the stage as well, I was like, what is happening? Like, how did you make to be on the stage with the guy and how does that feel? I mean he is a great artist, I mean crazy, but great

Clip 2


G: I mean shout out to R A the Rugged Man, he's become a really good friend, what's it called, no, it was cool, I mean, we're, we, like, we hit it off and we were chopping it with RA, Yella Brother knew who RA was and at that time RA broke his ankle, so he was

K: yeah, I saw him on the wheelchair

G: yeah, he was rapping on a wheelchair and he was like yo, I need someone to back me up, y'all down to back me up? And we were like alright and that's literally how it happened and then we went into a car and practiced the song for 20 minutes and he said alright we are up and I was like holy shit, okay, we are doing this

K: it was great show though

G: thank you


K: I mean, the vibe and all that plus you know, just him being on the stage it makes you feel like, you don't know what's gonna to happen the next minute, at least you know the songs, you know, show is like a different world especially with that guy

G: she did throw a few sticks in the crowd, so it's crazy

K: yeah, and I've got a question for you as well, cause I know that you have background with like jazz music and blues music in your family, so you sing as well, which is kind of, well a rare thing for an MC to do, so how did hip-hop came about then?


O: oh man, that's a great question, you did your research

K: I did yeah

O: You're killing the game, you know, growing up obviously my dad was a blues musician, so I grew up on the side of stages and you know, I think that blues and hip-hop are a lot alike, you know, it gives you really a space to talk about hardships and struggles and experiences, so when I was a teenager, or preteen, like 12-13 years old, you know, I needed an outlet that was more relevant to me, you know, and on the schoolyard everybody was freestyle battling each other and stuff like this, so I went from writing poems and songs on my guitar to being like, I think I want to step in the ring and, you know, in the year of 8th grade, like 12 years old, I just made the switch, I just put down my guitar and I said I'm going to step into the ring and start battling


O: only recently have I really started to dig deeper into the singing and stuff like that, because I really took like 10 years away from singing and, my girlfriend actually back home, she is a beautiful singer, she vocally produced my whole last album, she has actually been encouraging me to sing a bit more and stuff like that, I feel way more comfortable rapping, but the music I listen to the most is like R&B, you know and base and oldies and jazz, like, I just want to vibe out, like really heavy, so, she's like if you love it so much just start writing it


O: and so I have, it's really cool, it's the sensitive side

K: second confirmation, yeah, because, I mean, you must feel that it is a bit different when there is a girl on the stage and when there is a guy in the stage, obviously like, we know, how it is, like hip-hop, like it doesn't matter, like your age, race, nothing really, but I guess it might be so, do you get a sort of like a different vibration from people, like, do people feel like, uhm I'm not really sure how is that going to go?


O: totally, low expectations, well, now that, I mean obviously with Gavlyn, you know she's blown the fuck-up, especially, you know seeing her on stage, people already know what to expect from her, so it's different for me to be on this tour, where people know what to expect from a girl that's really good at rapping, but you know, playing for a crowd that doesn't, you know has no idea what's gonna come, you know it does come with low expectations, and when you can break those expectations and really just blow it out of the water, and make people surprised, I think that's one of the greatest feelings and accomplishments, so, (K: yeah) despite those low expectations, like, it's a challenge kind of, we have the opportunity to show somebody

G: yeah

O: something against the stereotype

K: girl power

O: exactly


G: like two girls on stage are already, like oh God

O: what are they doing?

K: and then it's like jaw dropping, oh like my God what's happening?

G: it always happens every time


K: that's good

G: so it's exciting

K: like, imagine that you would have all the money, all the talents in the world, would you still do what you

Clip 3


K: do, would you, would you want to try something else, something new? Did you ever thought about that?

G: I would probably do what I do and mesh it with a bunch of shit

K: okay

G if I was like, had all the extra talents that I wanted, I would be like cool I'm gonna to turn this rap music into something


K: so it would be like dancing, and singing and the DJ’ing at the same time

G: all of that, all of that, yep

K: okay, what about you O'Blimey?

O: it's crazy like, I feel like this has been a common question I've had in my head it's like, what would I do with all the funding in the world, you know

K: yeah

O: because, starting as like an independent artist, you know, not really having the money to start off with, you really see how vital it is to kind of make a budget and delegate, you know certain funds to certain areas, but thinking about what if I started with a bunch of money, where would I be right now, it's like crazy, you know, I could be DJ'ing, you know and I could have the voice box on stage, so I could be singing like crazy songs and like, you know, I would also probably like run a gym and have you know crazy fitness like programmes going on and stuff, you know, because I'm really into that sort of stuff, but like, just like the little sacrifices that you have to make as an independent artist running an independent career and when you think about not having to make those financial sacrifices, and everything just kind of being financially more easy, it's a difficult question to ask yourself, you know, where would I be if? But,


K: well, it's hard to imagine, so it's like, you know one way you lived, and then you know, the thing is that I think people also think about what they could have done, but actually people do have the money or something, or already are up there, they just think like like you people as well, because obviously, you do what you love, and they do what they do as well

O: totally

K: like, if there is money or connections or whatever, yeah,

O: I get it

K: I guess it's something that you can think about, it's good you guys would still do what you do which means like you are really dedicated, pretty dope actually, would you name like three things that you can't live without, whether it is like food, drink, music, I don't know people, places, anything, like three things


O: Wi-Fi

K: yes, that would be my one probably

G: music's definitely one of them too

O: girls, for me, personally

G: <laughs>


G: sure, okay I would say music, food and weed <laughs>

O: music food and weed <laughs>


O: so Wi-Fi, music, girls

K: alright, that's a nice combination as well, guys, so is there anything that we should be waiting for right now, do you have any more, obviously materials, you've got some video clips coming up, just shout out what you've got coming up


G: sure, I'm working on my brand-new project called make up for your breakup, it's like, it's probably the most mature sexual, feminine record I've ever done

O: it's super cool, it's crazy good

G: and I feel it's my best shit ever and other than that I do have a video clip coming out this thing called black chain cool lady, keep your eyes peeled for that and already have 2 records out: 'from the art' and 'modest confidence' go get that at Gavlyn.com


O: I am, I just finished a mix tape, shout out to Luke, who executive produced that, also executive producing her new mix tape which is gonna be incredible, I finished a mix tape called manifesto, which is m-n-f-s-t-o and there's gonna be a couple of new music videos coming out for that and a couple singles that I have got coming out, a song called Petaluma, which I, which actually is the only singing track that I have and you guys got to hear a little bit, so a video for Petaluma and a single for Petaluma coming out and yeah


K: that's great, make sure you guys check them out, so thanks again Gavlyn and O'Blimey

G: thank you for having us

O: shout out to say what

G: yeah, shout out say what

K: okay


Al-Aynati, Maamoun and Chorneyko, Katherine A. (2003) “Comparison of voice-automated transcription and human transcription in generating pathology reports”, Archives of pathology and laboratory medicine 127, no 6: 721-725.

Anasuya, M.A. and Katti, S.K. (2009) “Speech recognition by machine: A review”, International Journal of Computer Science and Information Security 6, no 3:181-205.

Arumí-Ribas, Marta and Romero-Fresco, Pablo (2008) “A practical proposal for the training of respeakers”, Journal of Specialised Translation 10:106-127.

Bailey, Julia (2008) "First steps in qualitative data analysis: transcribing", Family Practice 25, no 2: 127-131.

Bendazzoli, Claudio and Sandrelli, Annalisa (2007) “An approach to corpus-based interpreting studies: developing EPIC (European Parliament Interpreting Corpus” in Proceedings of the Marie Curie Euroconferences MuTra: Challenges of Multidimensional Translation, Sandra Nauert (ed), Saarbrücken, available online: http://www.euroconferences.info/proceedings/2005_Proceedings/2005_Bendazzoli_Sandrelli.pdf (last accessed 18 October 2016)

Bettinson, Mat (2013) The effect of respeaking on transcription accuracy. Honours thesis, unpublished. Melbourne: University of Melbourne.

Bucholtz, Mary (2000) “The politics of transcription”, Journal of Pragmatics 32: 1439–1465.

Cole, Alaistair (2015) Good Morning. Grade One. Language ideologies and multilingualism within primary education in rural Zambia, PhD diss, University of Edinburgh.

Davidson, Christina (2009) “Transcriptions: imperatives for qualitative research”, International Journal of Qualitative Methods, 8, no 2: 35-62.

Díaz-Cintas, Jorge and Orero, Pilar (2006) “Screen translation: voice-over” in Encyclopedia of languages, Keith Brown (ed), London: Elsevier, 473.

D’Arcangelo, Rossella and Cellini, Francesco (2013) “Metodi e tempi di una verbalizzazione - prove tecniche” in Numero monografico sul Respeaking, Specializzazione on-line, Carlo Eugeni and Luigi Zambelli (eds), 81-95.

Dumouchel, Pierre, Boulianne, Gilles and Brousseau, Julie (2011) “Measures for quality of closed captioning” in Audiovisual translation in close-up, Adriana Serban, Anna Matamala and Jean-Marc Lavour (eds), Bern, Peter Lang: 161-172.

Eugeni, Carlo (2003) “Respeaking the BBC News. A strategic analysis of Respeaking on the BBC”, The sign language translator and interpreter, 3, no 1: 29-68.

Eugeni, Carlo (2008a) La sottotitolazione in diretta TV. Analisi strategica del rispeakeraggio verbatim di BBC News, PhD diss, Università Degli Studi di Napoli Federico II.

Eugeni, Carlo (2008b) “A sociolinguistic approach to real-time subtitling: respeaking vs. shadowing and simultaneous interpreting” in English in international deaf communication, Cynthia J. Kellet Bidoli and Elana Ochse (eds), Bern, Peter Lang: 357-382.

Eugeni, Carlo and Mack, Gabriele (eds) (2006) New technologies in real time intralingual subtitling. Intralinea. Special issue: Respeaking.

Franco, Eliana, Matamala, Anna and Orero, Pilar (2000) Voice-over translation: an overview, Bern, Peter Lang.

Green, Judith, Franquiz, Maria, and Dixon, Carol (1997) “The myth of the objective transcript: Transcribing as a situated act”, TESOL Quarterly, 21, no 1: 172–176.

Hartman, Douglas J. (2015) “Enhancing and Customizing Laboratory Information Systems to Improve/Enhance Pathologist Workflow”, Surgical Pathology Clinics, 8, no 2: 137-43.

Joshi, Vivek, Narra, Vamsi, Joshi, Kailash, Lee, Kyootai, and Melson, David (2014) “PACS administrators' and radiologists' perspective on the importance of features for PACS selection”, Journal of Digital Imaging, 27, no 4: 486-95.

Lambourne, Andrew (2006) “Subtitle respeaking”, in New technologies in real time intralingual subtitling. Intralinea. Special issue, Carlo Eugeni and Gabriele Mack (eds), http://www.intralinea.org/specials/article/1686 (last accessed 01 March 2016)

MAA (Media Access Australia) (2014) Caption quality: international approaches to standards and measurement, Sydney, Media Access Australia http://www.mediaaccess.org.au/research-policy/white-papers/caption-quality-international-approaches-to-standards-and-measurement (last accessed 01 March 2016)

Mereghetti, Emiliano (2006) “La necessità dei sordi”, in New technologies in real time intralingual subtitling. Intralinea. Special issue, Carlo Eugeni and Gabriele Mack (eds), http://www.intralinea.org/specials/article/1697 (last accessed 01 March 2016)

Muzii, Luigi (2006) “Respeaking e localizzazione”, in New technologies in real time intralingual subtitling. Intralinea. Special issue, Carlo Eugeni and Gabriele Mack (eds), http://www.intralinea.org/specials/article/1688 (last accessed 01 March 2016)

Ochs, Elinor (1979) “Transcription as theory”, in Developmental pragmatics, Elionor Ochs and Bambi Schiefflin (eds), New York: Academic, 43-72.

Ofcom (Office of Communications) (2013) Measuring the quality of live subtitling: Statement, London, Ofcom. http://stakeholders.ofcom.org.uk/binaries/consultations/subtitling/statement/qos-statement.pdf (last accessed 01 March 2016)

Oliver, Daniel G., Serovich, Julianne M., and Mason, Tina L. (2005) “Constraints and opportunities with interview transcription: Towards reflection in qualitative research”, Social Forces, 84, no 2: 1273–1289.

Pollak, Alexander (2008) “Analyzing TV Documentaries”, in Qualitative Discourse Analysis in the Social Sciences, Ruth Wodak and Micha Krzyanowski (eds), Houndmills: Basingstoke, Hampshire, 77-95.

Prazac, Ales, Loose, Zdenek, Trmal, Jan, et al. (2012) “Novel approach to live captioning through re-speaking: tailoring speech recognition to re-speaker’s needs”, InterSpeech 2012: 1372-1375.

Ramondelli, Fausto (2006) “La sottotitolazione in diretta con la stenotipia” in New technologies in real time intralingual subtitling. Intralinea. Special issue, Carlo Eugeni and Gabriele Mack (eds), http://www.intralinea.org/specials/article/1694 (last accessed 01 March 2016)

Romero-Fresco, Pablo (2010) “Standing on quicksand: viewers’ comprehension and reading patterns of respoken subtitles for the news”, in New insights into audiovisual translation and media accessibility, Jorge Díaz-Cintas, Anna Matamala and Josélia Neves (eds), Amsterdam, Rodopi, 175-195.

Romero-Fresco, Pablo (2011) Subtitling through speech recognition: Respeaking, Manchester, Routledge.

Romero-Fresco, Pablo (2012a) Respeaking in translator training curricula: present and future prospects, The Interpreter and Translator Trainer 6, no 1:91-112.

Romero-Fresco, Pablo (2012b) “Quality in live subtitling: the reception of respoken subtitles in the UK”, in Audiovisual translation and media accessibility at the crossroads, Aline Remael, Pilar Orero and Mary Carroll (eds), Amsterdam, Rodopi, 111-133.

Romero-Fresco, Pablo (2013) “Accessible filmmaking: Joining the dots between audiovisual translation, accessibility and filmmaking”, Jostrans - The Journal of Specialised Translation 20: 201–223.

Romero-Fresco, Pablo (2016) “Accessing communication: The quality of live subtitles in the UK”, Language & Communication 49: 56–69.

Romero-Fresco, Pablo and Martínez, Juan (2015) “Accuracy rate in live subtitling – the NER model”, in Audiovisual translation in a global context. Mapping an ever-changing landscape, Rocío Baños Piñeiro, and Jorge Díaz-Cintas (eds), London: Palgrave Macmillan, 28-50.

Singh, Meenakshi, Pal, Timothy (2011) “Voice recognition technology implementation in surgical pathology: advantages and limitations”, Archives of Pathology and Laboratory Medicine, 135, no 11: 1476-81.

Sohn, Shara D. (2004) Court reporting: can it keep up with technology or will it be replaced by voice recognition or electronic recording? Honours thesis, Southern Illinois University. http://opensiuc.lib.siu.edu/cgi/viewcontent.cgi?article=1264&context=uhp_theses (last accessed 01 March 2016)

Sperber, Matthias, Neubig, Graham, Fügen, Christian, et al. (2013) “Efficient speech transcription through respeaking”, InterSpeech 2013: 1087-1091.

Trivulzio, Fausto (2006) “Natura non facit saltus”, in New technologies in real time intralingual subtitling. Intralinea. Special issue., Carlo Eugeni and Gabriele Mack (eds), http://www.intralinea.org/specials/article/1690 (last accessed 01 March 2016)

Van Waes, Luuk, Leijten, Mariëlle and Remael, Aline (2013) “Live subtitling with speech recognition. Causes and consequences of text reduction”, Across languages and cultures, 14, no 1:15-46.

Wald, Mike (2009) “Real-time speech recognition subtitling in education”, 2nd International Seminar on Real-Time Intralingual Subtitling. Barcelona, UAB.

About the author(s)

Anna Matamala, BA in Translation (UAB) and PhD in Applied Linguistics (UPF), is since 2009 a senior tenured lecturer at Universitat Autònoma de Barcelona. A member of the research group Transmedia Catalonia (http://grupsderecerca.uab.cat/transmedia/), Anna Matamala has participated (DTV4ALL, ADLAB, HBB4LL) and led (AVT-LP, ALST, VIW) funded projects on audiovisual translation and media accessibility. She is currently leading the project NEA and is involved in the European projects ACT, ADLAB PRO and ImAC. She has taken an active role in the organisation of scientific events (Media for All, ARSAD), and has published extensively in international journals such as Meta, The Translator, Perspectives, Babel, Translation Studies, among others. She is currently involved in standardisation work at ISO. Website: gent.uab.cat/amatamala

Pablo Romero Fresco is a Ramón y Cajal grantholder at Universidade de Vigo (Spain) and Honorary Professor of Translation and Filmmaking at the University of Roehampton (London, UK). He is the author of the book Subtitling through Speech Recognition: Respeaking (Routledge) and the editor of The Reception of Subtitles for the Deaf and Hard of Hearing in Europe (Peter Lang). He has collaborated with several governments, universities, companies and user associations around the world to introduce and improve access to live events for people with hearing loss around the world. He is the leader of the EU-funded projects “MAP: Media Accessibility Platform” and “ILSA: Interlingual Live Subtitling for Access” and of the international research hub OGAM (Galician Observatory for Media Access). Pablo is also a filmmaker and has developed the notion of accessible filmmaking, the integration of translation and accessibility as part of the filmmaking process. His first documentary, Joining the Dots (2012), was used by Netflix as well as schools around Europe to raise awareness about audio description.

Lukasz Daniluk completed an MA in Audiovisual Translation with a focus on video game localization at the University of Roehampton. He is a freelance professional within the film and TV industry with main interests in video editing, translation and transcription. He is currently enrolled on a doctoral programme in which he is conducting an ethnographic study of rap and hip-hop culture in the Polish diaspora in Britain as well as Polish Hip-Hop culture in Poland. The study involves recording video interviews with various Polish Hip-Hop artists as part of the data collection and it utilises re-speaking as a means of data transcription. As a multi-theoretical study it explores identity, nationality, authenticity, cultural flows, transnationalism and linguistic globalisation.

Email: [please login or register to view author's email address]

©inTRAlinea & A. Matamala, P. Romero-Fresco & L. Daniluk (2017).
"The Use of Respeaking for the Transcription of Non-Fictional Genres An Exploratory Study"
inTRAlinea Volumes
Edited by: {specials_editors_volumes}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2262

Traduire à voix haute

la traduction dictée interactive comme solution ergonomique

By Julián Zapata & Elizabeth C. Saint (University of Ottawa, Canada)

Abstract & Keywords


Voice recognition (VR) technology has greatly evolved since computer tools first came into our lives. It has been increasingly adopted by translators, who turn to the technology to minimise or prevent certain physical or mental health problems, among other reasons. VR has also helped bring translation dictation back into the profession. As shown in the first study presented in this article, the ergonomic advantages of using VR are genuinely valued by translators. However, we argue that VR could be improved by combining it with the various input modes offered by newer multimodal and mobile interfaces, which offer many advantages. We therefore advocate moving beyond translation dictation to develop an interactive translation dictation (ITD) environment, through which the professional translator's workstation would become a natural extension of his or her capabilities. Our second study presents the results of an experiment designed to assess the ergonomic gains of using a prototypical ITD environment.


Les technologies de reconnaissance vocale (RV) se sont largement développées depuis l’arrivée des outils informatiques dans notre vie. De plus en plus employée par les traducteurs pour, entre autres, réduire ou prévenir certains problèmes de santé physiques ou psychologiques, la RV a aussi contribué au retour de la traduction dictée chez ces professionnels. Comme la première étude présentée dans cet article le montre, les avantages ergonomiques de l’utilisation de la RV sont véritablement appréciés par les traducteurs. Nous avançons toutefois que la RV gagnerait à être combinée aux multiples atouts que présentent les interfaces multimodales des outils technologiques émergents et mobiles. Plus que de traduction dictée, il s’agirait alors d’offrir un espace de traduction dictée interactive (TDI), par lequel la plateforme de travail du traducteur professionnel deviendrait un prolongement de ses compétences. Notre seconde étude présente les résultats d’une expérience d’utilisation d’un prototype d’environnement de TDI visant à en évaluer sa valeur ajoutée en matière d’ergonomie.

Keywords: voice recognition, multimodal interaction, interactive translation dictation, translation technology, ergonomics, reconnaissance vocale, interaction multimodale, traduction dictée interactive, traductique, ergonomie

©inTRAlinea & Julián Zapata & Elizabeth C. Saint (2017).
"Traduire à voix haute la traduction dictée interactive comme solution ergonomique"
inTRAlinea Volumes
Edited by: {specials_editors_volumes}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2261


Depuis les années 1950 et l’arrivée de l’ordinateur, les chercheurs[1] de l’informatique travaillent à développer des moyens de traiter automatiquement le langage humain. Leurs travaux ont donné lieu à la conception de divers systèmes employant les technologies de reconnaissance et de synthèse vocales, tels que les agents conversationnels Alexa, Cortana ou Siri. Ces deux technologies permettent à la machine de « reconnaître » automatiquement ce que dit une personne (reconnaissance vocale) et de générer automatiquement de la parole en simulant la voix humaine (synthèse vocale). La première s’est améliorée plus lentement que la seconde, mais on constate que les technologies de reconnaissance vocale (RV) arrivent enfin à maturité et qu’elles sont maintenant employées avec succès dans des domaines aussi divers que les services à la clientèle, le soutien technique téléphonique et en ligne, les systèmes de navigation des véhicules ou encore les systèmes d’exploitation des ordinateurs personnels, tablettes ou téléphones intelligents (Jurafsky et Martin 2009 : 285). Ces avancées coïncident avec le développement des systèmes informatiques multimodaux qui rendent l’interaction avec la machine plus naturelle et plus ergonomique[2].

Il n’est pas étonnant que la traduction profite, elle aussi, de ces progrès technologiques; d’autant que l’idée de jumeler la RV et la traduction dictée (Brousseau et coll. 1995; Brown et coll. 1994; Dymetman et coll. 1994) remonte aux années 1990. Toutefois, les imperfections de la RV à cette époque et son incapacité à automatiser des tâches langagières essentielles, comme la transcription de la dictée, ont contribué à ce que la recherche se tourne plutôt vers la conception d’outils de traductique, capables de prendre en charge des tâches linguistiques périphériques et d’améliorer la performance et l’efficacité des traducteurs (Bowker 2002). Ainsi, ce n’est réellement qu’au début des années 2000 que la RV a commencé à faire partie de l’outillage des traducteurs et, rapidement, elle a présenté de nombreux avantages, notamment ergonomiques. La valeur ajoutée de la RV dans la pratique professionnelle du traducteur justifie que l’on vise son intégration au sein d’une interaction multimodale avec la machine. C’est du moins la proposition que fait cet article en présentant la traduction dictée interactive (TDI), concept innovant introduit par Zapata (2012; 2016) pour l’amélioration des conditions de travail des traducteurs professionnels et de leur productivité.

Nous commencerons par faire un rapide retour historique sur l’intégration de la RV dans la pratique de la traduction et nous présenterons brièvement les résultats d’une étude qui révèlent les avantages que des traducteurs disent tirer de la RV lorsqu’ils l’utilisent pour traduire. Ceci nous amènera à exposer les caractéristiques de l’interaction multimodale et du concept de TDI. Notre article se conclura sur quelques-uns des résultats d’une expérience menée auprès de 14 traducteurs professionnels qui ont traduit à partir d’un poste de travail traditionnel (avec clavier et souris) et avec un environnement de TDI. Nous limiterons notre discussion aux données qualitatives issues des entretiens semi-dirigés avec les participants, durant lesquels ils se sont exprimés sur l’ergonomie des deux environnements de travail.

La recherche d’une traductique ergonomique

La reconnaissance vocale (RV)[3] se définit comme une « [t]echnologie ayant pour but de permettre à un ordinateur de reconnaître les signaux émis par la voix humaine en vue de la transformer en données numériques » (Office québécois de la langue française 2007). Elle se situe dans le domaine du traitement automatique des langues naturelles (TALN), c’est-à-dire tout ce qui unit l’informatique au langage humain. Dès les prémices du TALN, dans les années 1950, l’un des objectifs était de faire de l’ordinateur un agent conversationnel, un interlocuteur de l’être humain. Toutefois, au début des années 1990, on était encore loin d’arriver à cet idéal et la RV éprouvait encore de nombreuses difficultés à analyser la parole humaine dans toutes ses variétés (accent, intonation, sonorité, vitesse, timbre de la voix, etc.). Dans le courant de cette décennie, de nombreux laboratoires d’entreprises commerciales ont largement investi pour développer des systèmes de RV fonctionnels, pouvant diriger des machines pour des tâches bien spécifiques. Ces avancées, satisfaisantes pour les uns (Rabiner 1997), mais frustrantes pour d’autres (Lippmann 1997), ont suscité un intérêt soudain pour le TALN dans le domaine de la traduction (Brousseau et coll. 1995; Brown et coll.1994; Dymetman et coll. 1994). Toutefois, la RV d’alors n’étant pas encore assez performante pour répondre aux besoins de cette profession, c’est le traitement de tâches linguistiques périphériques qui a été privilégié et qui a donné naissance aux outils de traductique (Bowker 2002). Il a fallu attendre le début des années 2000 et l’arrivée de logiciels de RV commerciaux plus performants pour que cette technologie refasse son apparition dans la boite à outils de quelques traducteurs (voir, par exemple, Benis 2002; Seaman 2002; Stroman 2002).

Les chercheurs recommencent alors à s’intéresser à l’intégration de la RV dans la pratique professionnelle de la traduction, notamment pour tester si elle aide le traducteur à gagner en productivité (Désilets et coll. 2008; Dragsted, Hansen et Selsøe Sørensen 2009; Reddy et Rose 2010; Vidal et coll. 2006). Le succès des systèmes de RV dans d’autres domaines, la performance de traitement accrue des ordinateurs intégrant la RV et le besoin urgent de concevoir des outils de traductique ergonomiques, c’est-à-dire qui tiennent compte du facteur humain (O’Brien 2012), ont éveillé un réel intérêt chez les chercheurs en traductologie pour les technologies de RV. Ainsi, certains travaux se sont intéressés à l’usage de la RV dans la formation des traducteurs (Mees et coll. 2013) ou des post-éditeurs (Mesa-Lao 2014), ainsi que dans la pratique du sous-titrage (Romero-Fresco 2011). D’autres ont testé la performance de la RV pour la dictée sur différentes interfaces informatiques (Zapata et Kirkedal 2015) ou pour la post-édition de traductions automatiques (Garcia-Martinez et coll. 2014; Torres-Hostench et coll. 2017; Zapata, Castilho et Moorkens 2017). Finalement, quelques-unes se sont penchées sur la perception qu’ont les traducteurs professionnels des avantages de la RV, notamment, pour leur productivité et pour la qualité de leurs traductions (Ciobanu 2014; 2016).

Ces recherches confirment qu’à l’instar du reste des utilisateurs des technologies, la dépendance du traducteur aux systèmes informatiques est en constante progression. Face au besoin palpable d’interfaces spécifiquement ajustées au processus traductif et aux besoins ergonomiques des traducteurs (Carl, Dragsted et Jakobsen 2011; Taravella et Villeneuve 2011; 2013), un appel à de plus amples études expérimentales a été lancé :

More experimental studies of translator-tool interaction could be carried out using formal usability research methods such as screen recording, eye tracking, and observation, the results of which could then be used by translation technology developers to improve the specifications of tools for the benefit of translators and, ultimately, the end users of those translations. (O’Brien 2012 : 116–7)

Étant donné qu’une technologie trop rigide entrave le travail qu’elle est censée appuyer (Karamanis, Luz et Doherty 2011 : 49), les traductologues sont de plus en plus conscients, qu’en cette ère de traductique, il leur faudra accorder une plus grande place à l’étude du facteur humain (O’Brien 2012). Cela permettra alors de concevoir des interfaces logicielles, plus conviviales et mieux adaptées, à la fois aux besoins ergonomiques et documentaires des traducteurs et à la réalité d’un marché mondial de la traduction qui s’élargit et évolue rapidement (Barabé 2013; Taravella et Villeneuve 2013). À ce propos, la première étude que nous exposons ici vise à extraire les avantages ergonomiques et documentaires que ressentent les traducteurs qui emploient la RV dans leur pratique.

Étude 1 : Les avantages de l’utilisation de la RV selon les traducteurs (ergonomie, productivité, qualité)

Le corpus que nous avons analysé est extrait de trois portails en ligne destinés aux traducteurs et se compose de commentaires publiés par ces derniers, entre janvier 2003 et mars 2012 (fin de la collecte de données), au sujet de la RV. Malgré sa petite taille (9 500 mots), notre corpus a l’avantage de présenter des propos spontanés de traducteurs intéressés par la traduction avec RV. Nous présentons les résultats de façon très globale, mais invitons le lecteur à se référer au travail de Zapata (2012) pour accéder à de plus amples détails.

Les raisons ergonomiques sont les premières évoquées par les traducteurs pour justifier leur utilisation de la RV dans leur pratique professionnelle et elles précèdent celles liées à leurs besoins de gagner en productivité ou en qualité. Ainsi, plusieurs indiquent adopter la RV pour deux raisons : de façon préventive, pour diminuer le stress ou éviter des maladies chroniques, comme des maladies musculo-squelettiques (ex. : le syndrome du canal carpien), des maux de dos, aux épaules, aux poignets ou des problèmes oculaires; ou comme mesure réparatrice après avoir contracté ces troubles. En règle générale, les traducteurs se servent de logiciels de RV pour dicter leur traduction et certains y voient, là aussi, des avantages pour leur santé et leur confort personnel. En effet, la dictée leur permet de s’éloigner de leur écran d’ordinateur et de rester productifs sans avoir à rester assis pendant plusieurs heures.

Outre les bienfaits de la mobilité pour la santé, quelques traducteurs indiquent que le fait de pouvoir utiliser un casque d’écoute sans fil ou de s’éloigner de leur poste de travail leur permet de travailler de façon plus décontractée, tout en garantissant la production d’une traduction de qualité. En effet, après avoir constaté que leur volume de travail dépassait leur capacité de saisie à l’ordinateur, certains traducteurs disent s’être tournés vers la RV et avoir atteint une productivité jusqu’à six fois plus élevée en dictant, plutôt qu’en employant un clavier. Un facteur contributeur évoqué est que l’emploi de la RV faciliterait la concentration du traducteur sur la tâche de traduction uniquement, ce qui, en retour, lui permettrait de soigner la qualité du produit final et le rendrait aussi plus créatif. Une poignée de traducteurs mentionnent regretter que l’utilisation du clavier ait un impact si nuisible sur leur créativité et leur compétence de traduction à proprement parler. En somme, selon eux, l’évaluation de la compétence d’un traducteur ne devrait pas avoir à dépendre autant de sa compétence à saisir (rapidement) sa traduction au clavier d’ordinateur.

Alors que les traducteurs emploient les systèmes de RV pour dicter leurs traductions, peu d’entre eux les utilisent pour envoyer des commandes à l’ordinateur ou pour fureter dans des dossiers ou dans le Web. Pourtant, ceux qui y recourent les trouvent particulièrement utiles pour corriger des erreurs de reconnaissance, surtout qu’en le faisant à voix haute, le logiciel de RV apprend des erreurs et devient du même coup plus performant. Quelques traducteurs démontrent un intérêt à envoyer des commandes à divers logiciels de traductique installés sur leur ordinateur, notamment les gestionnaires de mémoire de traduction, mais reconnaissent qu’ils ne sont pas tous compatibles avec la RV. Finalement, des traducteurs développent une interaction que l’on peut considérer de multimodale (Oviatt 2012) en alternant entre la RV, le clavier et la souris pour dicter leurs traductions. Cette forme d’interaction leur permettrait de contrer les problèmes de reconnaissance ou d’incompatibilité des logiciels de traductique, par exemple. Ces traducteurs ont donc su trouver un équilibre dans cette alternance et tirer profit de la RV, là où elle fonctionne sans erreur ou presque.

L’utilisation de la RV apporterait donc une nouvelle dimension au flux de travail habituel pour les traducteurs. En assurant son adaptation optimale aux outils de traductique, la traduction dictée (TD) pourrait devenir une TD interactive (TDI) qui amènerait le traducteur à interagir avec la machine de façon multimodale, ergonomique, intuitive ou, tout simplement, « humaine » (ibid. : 428).

L’interaction multimodale, modèle ergonomique

Les systèmes multimodaux visent à reconnaître les comportements humains et l’utilisation naturelle du langage, afin de parvenir à interagir le plus intuitivement possible avec la machine. Oviatt (ibid. : 405) les définit ainsi :

Multimodal systems process two or more combined user input modes – such as speech, pen, touch, manual gestures, gaze, and head and body movements – in a coordinated manner with multimedia system output. (…) This new class of interfaces aims to recognize naturally occurring forms of human language and behavior, which incorporate at least one recognition-based technology (e.g. speech, pen, vision).

En cela, les systèmes multimodaux remettent en question la pertinence des interfaces graphiques traditionnelles telles que les ordinateurs de bureau, qui font essentiellement usage du clavier et de la souris, en explorant d’autres modes de saisie et de sortie (Girol et coll. 2012; Williamson, Crossan et Brewster 2011).

Les avantages de l’interaction multimodale avec la machine sont nombreux, allant d’une meilleure performance (Cohen et coll. 1998) à une flexibilité d’utilisation accrue (usage en extérieur, en déplacement, etc.) (Oviatt et Cohen 2000). C’est aussi ce que révèle une étude de cas menée par Schalkwick et coll. (2010), de l’entreprise Google, sur l’un des exemples actuels les plus concrets et universels d’interface multimodale : le téléphone intelligent. En testant l’usage des commandes vocales pour faire des recherches sur le Web et dans le système d’exploitation du téléphone, les chercheurs ont souligné les nombreux avantages liés à ce type d’interaction, tout en reconnaissant les nombreux défis liés à sa conception. Ainsi, en autorisant la saisie par le toucher et la RV, le téléphone intelligent confirme les gains en mobilité, flexibilité et performance qu’apporte la multimodalité et fournit un modèle d’interaction plus naturelle entre l’humain et la machine. En outre, la polyvalence du système permet à l’utilisateur d’employer simultanément des modes de saisie passifs (clavier, souris) et actifs (voix, doigt, stylet) (Shneiderman 1998 : 318) ou d’alterner entre eux plus aisément, ce qui semble réduire le niveau de frustration des utilisateurs envers la machine, éviter les erreurs (Oviatt et van Gent 1996) et alléger la surcharge cognitive face à des tâches complexes (Oviatt, Coulston et Lunsford 2004).

Même du point de vue de la conception des interfaces, les systèmes multimodaux se sont montrés avantageux, puisque capables de servir des utilisateurs aux profils divers, sans avoir à faire de nombreux compromis pour atteindre l’« utilisateur moyen » (si tant est que celui-ci existe). En somme, les interfaces multimodales ont non seulement le potentiel de fonctionner plus efficacement que les interfaces monomodales, mais elles le font aussi d’une façon plus stable dans des contextes divers et pour des utilisateurs aux profils variés. Les différences comportementales ont d’ailleurs inspiré et guidé un bon nombre des recherches sur la multimodalité. Ainsi, une série d’études empiriques menées par Oviatt, Cohen et Wang (1994) mettent l’accent sur les défis des interfaces multimodales à traiter la variation linguistique des utilisateurs, à l’oral comme à l’écrit. Une autre recherche s’est intéressée aux raisons qui mènent les utilisateurs à interagir différemment avec l’interface multimodale et a tenté de vérifier si elles sont liées à d’autres caractéristiques comportementales (Oviatt, Lunsford et Coulston 2005). En traduction, où les questions de l’ergonomie gagnent en importance, la recherche sur la multimodalité se penche, par exemple, sur les usages de la voix, du stylet ou de l’écran tactile. Pour notre part, nous explorons une approche multimodale de la TD.

Au-delà de la dictée : la TDI

La TDI se définit comme une « technique traductive qui implique l’interaction avec des interfaces multimodales tactiles et vocales, tout au long du processus de traduction, soit la préparation, la production et la révision » (Zapata 2016 : 86, notre traduction). Elle représente une nette évolution des techniques de traduction courantes, comme la traduction saisie à l’ordinateur et la TD, que cette dernière soit traditionnelle avec dictaphone ou qu’elle emploie la RV.

Comme le montre la Figure 1, la TDI constitue une forme de traduction à vue (TAV), technique de traduction à voix haute, employée depuis des décennies dans la formation et la pratique de l’interprétation, mais aussi dans celles de la traduction (Jiménez Ivars et Hurtado Albir 2003).


Figure 1. Les divers types de TAV

Avant de présenter la TDI plus en détail, et pour mieux comprendre en quoi elle diffère des méthodes de traduction courantes, nous prendrons quelques instants pour décrire ces dernières (traduction saisie à l’ordinateur et autres traductions dictées). Nous excluons de la discussion la traduction dictée directement à un copiste, cette pratique étant, semble-t-il et selon les dires de traducteurs professionnels avec qui nous avons pu nous entretenir, maintenant rare et déconseillée pour des questions de rendement et de rentabilité. Ce qui différencie ces approches réside dans la manière dont la traduction sera transcrite (Figure 2).


Figure 2. Les modes de transcription des TD, autres que la TDI

Ainsi, la traduction saisie à l’ordinateur, c’est-à-dire manuellement par l’usage d’un clavier, s’avère moins efficace que les TD pour la concentration du traducteur sur le texte à traduire. En effet, de récentes études (Dragsted et Hansen 2009; Dragsted, Hansen et Selsøe Sørensen 2009; Mees et coll. 2013) ont montré que le fait de produire sa traduction oralement, à la vitesse de la parole, aide le traducteur à se concentrer sur le texte de départ et l’autorise à laisser les questions de transcription et de révision pour plus tard. Cela est particulièrement le cas de la TD au dictaphone et de la TD au système de RV qui seront transcrites de manière consécutive (voir Figure 2). Toutefois, si la TD au dictaphone profite à la concentration sur le texte source, elle présente l’inconvénient d’ajouter la tâche de transcription au traducteur ou l’embauche de personnes complémentaires, à moins, bien sûr, que le client n’accepte de se charger lui-même de la transcription. En ce qui concerne la TD au système de RV avec transcription simultanée, Dragsted, Hansen et Selsøe Sørensen (2009 : 313) ont indiqué qu’elle présentait de plus grands défis d’attention pour le traducteur. De fait, voyant sa traduction s’afficher à l’écran au fur et à mesure qu’il la dicte, il aurait tendance à ne pas résister à la tentation de corriger les erreurs de transcription qui peuvent apparaître. La TD au système de RV avec transcription consécutive vient, sans aucun doute, régler ce problème puisque le traducteur soumet l’ensemble de sa dictée enregistrée à un logiciel de RV qui, à son tour, prend en charge la transcription (Wu 2008). Par contre, cette approche requiert du traducteur qu’il soit équipé d’une variété d’appareils et de logiciels, tels qu’un ordinateur personnel puissant, un dictaphone numérique, un système de RV (unilingue et monomodal) et un logiciel de traitement de textes. Enfin, une autre approche a été envisagée par certains organismes[4] : le travail en équipe de traducteurs dictant des traductions à l’aide de dictaphones numériques et de copistes transcrivant les enregistrements à l’aide de logiciels de RV. D’ailleurs, certains marchands offrent des solutions logicielles permettant ce type de travail collaboratif.

En somme, les TD auront présenté de nombreux avantages sur la traduction saisie à l’ordinateur, mais elles comportent encore de nombreux inconvénients que la TDI pourrait, croyons-nous, estomper en offrant une interaction multimodale et ergonomique entre les traducteurs et la machine (voir Figure 3).


Figure 3. Description de la TDI dans le processus traductif

En effet, en permettant au traducteur d’utiliser le système de RV pour faire plus que de la TD (par exemple, de la recherche documentaire et terminologique dans les corpus ou bases de données terminologiques) et d’y ajouter l’usage du stylet ou du toucher quand le besoin se fait sentir, la TDI a le potentiel de devenir « un prolongement naturel du processus de travail habituel du traducteur professionnel » (Zapata et Quirion 2016 : 539). Cela vaut d’autant plus que les interfaces multimodales commerciales à notre disposition (tablettes, ordinateurs à écran tactile) n’ont jamais été aussi robustes et qu’elles ne vont qu’en s’améliorant. La deuxième étude que nous exposons ici vise précisément à tester cette hypothèse.

Étude 2 : Les avantages de l’utilisation d’un environnement de TDI selon les traducteurs

Lors de cette expérience, 14 traducteurs professionnels qui travaillent de l’anglais vers le français ont eu à effectuer une traduction en interaction soit avec un poste de travail traditionnel (avec clavier et souris) soit avec un environnement de TDI (Figure 4). Ce dernier consistait en deux interfaces multimodales commerciales, à savoir une tablette et un ordinateur à écran tactile connectés à l’Internet et équipés d’une variété de logiciels (suite bureautique, outils de traductique, système de RV, clavier virtuel, etc.), ainsi qu’un stylet.


Figure 4. Prototype d’un environnement de TDI

Des données quantitatives et qualitatives ont été relevées au cours de l’expérience, les premières touchant à des aspects tels que la vitesse de frappe au clavier et de lecture à voix haute des traducteurs, l’exactitude de la transcription et l’usage de la souris. Les données qualitatives ont été récoltées lors d’entretiens semi-dirigés avec les traducteurs. L’un des objectifs de ces entretiens était d’extraire les perceptions des traducteurs sur leur utilisation d’un environnement de TDI pour traduire un texte, comparé à leur environnement de travail habituel (ordinateur, clavier et souris). L’hypothèse sous-jacente à l’expérience était que l’environnement de TDI améliorerait l’expérience de travail du traducteur grâce à une interaction plus conviviale avec la machine. Les résultats que nous présentons ici ne concernent que les aspects ergonomiques évoqués par les traducteurs[5], puisque ces derniers ont été un sujet fort récurrent dans les conversations.

Tout d’abord, dans leur configuration de travail habituelle, les traducteurs sont dotés d’un ordinateur, d’un écran, d’une souris et d’un clavier. Toutefois, pour des questions d’amélioration de la productivité et de prévention de maladies, les configurations de travail de quelques traducteurs ont été modifiées. Par exemple, certains travaillent avec deux écrans qu’ils utilisent respectivement pour afficher le texte source et le texte cible ou pour faire apparaître les deux textes sur un écran et utiliser le second pour la recherche documentaire. Deux participants ont, eux, reçu un clavier et une souris ergonomiques qui leur permettent de placer leurs mains et poignets dans des positions « optimales ». L’un d’entre eux précise aussi avoir reçu l’aide d’un spécialiste en ergonomie pour obtenir une chaise ajustée à sa taille et un placement en angle de ses deux écrans.

Hormis une personne qui se dit entièrement satisfaite de son environnement de travail actuel et indique n’avoir « jamais ressenti [de] fatigue au travail », les autres participants admettent vivre des douleurs physiques, particulièrement aux poignets, dues à la position assise devant un écran qu’ils doivent maintenir pendant de longues heures. Parmi les moyens que déploient les traducteurs pour pallier ces problèmes se trouve la variété des tâches qui les amènent à ne pas toujours faire les mêmes mouvements. Ainsi, deux traducteurs indiquent imprimer le premier jet de leur traduction afin d’appliquer les révisions avec un crayon, ce qui les aiderait aussi à mieux repérer les erreurs. Un autre traducteur choisit aussi d’imprimer son texte source et de l’annoter au crayon de toutes les informations qui lui seront utiles pour traduire. Finalement, le kinésithérapeute d’un des participants lui aurait conseillé de prendre des pauses et de se déplacer régulièrement, mais cela semble difficile à mettre en pratique en raison des délais et des objectifs de productivité journaliers à respecter.

On le voit donc, tous les ajustements sont dus au fait que les traducteurs sont, en grande majorité, soucieux et conscients des implications que la configuration de leur environnement de travail peut avoir sur leur santé. Le participant qui a reçu l’appui du spécialiste en ergonomie dit même avoir voulu recevoir une évaluation dès le début de son activité professionnelle, car il sait que les traducteurs sont sujets aux tendinites du fait qu’ils passent, selon lui, 95 % de leur temps à taper à l’ordinateur. Dans le contexte qui est le sien, il juge son environnement de travail ergonomique et optimal. En revanche, il reconnaît qu’avec un matériel comme celui du prototype d’un environnement de TDI, les traducteurs seraient « dans une classe à part ».

Cela nous amène aux commentaires des participants au sujet de leur utilisation de cet environnement de TDI. Sans grande surprise, les traducteurs ont apprécié la possibilité de travailler tout en pouvant se déplacer et ont évoqué les bienfaits de cette option pour les mêmes aspects de santé qui les préoccupent dans leur contexte de travail actuel. Notons que même le traducteur qui indiquait être entièrement satisfait de son environnement professionnel a reconnu aimer pouvoir bouger en travaillant et a admis qu’il aimerait avoir accès aux deux environnements de travail. L’augmentation de la mobilité apportée par l’interface multimodale de TDI a aussi été soulignée par une personne qui a vu un avantage à pouvoir passer régulièrement d’une interface à une autre et d’un appareil à un autre. Selon elle, cela force aussi le traducteur à bouger du fait qu’il utilise le stylet, les doigts ou la tablette.

Selon un participant, l’ergonomie de l’environnement de TDI pourrait avoir un impact positif sur la qualité de la traduction puisqu’il favoriserait la concentration optimale du traducteur et préserverait son niveau d’énergie au cours de la journée. Ce constat de qualité améliorée du produit de la traduction est partagé par la majorité des traducteurs. Seules deux personnes pensent différemment, l’une ayant démontré un manque d’enthousiasme généralisé face à l’expérience et une autre doutant que la qualité de la traduction puisse être meilleure que celle obtenue à partir d’un environnement de travail traditionnel. Certains participants jugent que l’environnement de TDI diminue le niveau de stress en augmentant le nombre de mots traduits par minute et, du même coup, en accordant plus de temps à la préparation et à la recherche documentaire, ainsi qu’à la révision du texte cible. « Ce sont des choses qu’on n’a presque plus le temps de faire parce que les échéances sont trop serrées », indique l’un d’entre eux. Ce gain de temps se ressentira donc aussi sur la qualité du produit fini.

Toutefois, le gain de temps reste une estimation de la part des traducteurs et il n’a pas été ressenti lors de l’expérience. En effet, la plupart d’entre eux ont signalé que pour tirer réellement profit de l’environnement multimodal, il leur faudrait plusieurs heures de pratique, ne serait-ce que pour entraîner le système de RV à reconnaître les caractéristiques langagières de son utilisateur : « Côté productivité, je pense que ça ferait un creux avant de remonter. Il y a beaucoup de choses à intégrer. » Néanmoins, la majorité ayant pu observer les avantages de la dictée comparée à la saisie au clavier d’ordinateur, les traducteurs de notre étude tendent à voir le potentiel d’un tel environnement de TDI.


Dans cet article, nous avons tenté de mettre l’accent sur le potentiel ergonomique qu’offrent les technologies de reconnaissance vocale (RV) et les interfaces multimodales pour la traduction professionnelle. Deux études explorant l’usage de la RV et de la traduction dictée interactive (TDI) ont été présentées et nous avons discuté leurs résultats en examinant les avantages de la RV et de la TDI pour l’amélioration et le maintien du bien-être physique et psychologique des traducteurs.

L’ergonomie va bien au-delà de la simple fabrication de chaises de bureau confortables ou de claviers d’ordinateur courbés pour épouser la forme des poignets; c’est pourquoi, depuis longtemps, les études sur le sujet intègrent des questions touchant aux impacts physiques et psychologiques des technologies sur les êtres humains et sur leurs activités. C’est le cas, par exemple, des recherches considérant les effets de l’usage des ordinateurs sur l’acte d’écrire et sur la pensée créative qui y est associée (Daiute 1985). Ces études, souvent initiées suite à des plaintes d’utilisateurs, aident à améliorer les outils, les systèmes ou les logiciels de façon à ce qu’ils répondent aux besoins ergonomiques des personnes qui les emploient. Cette approche permettrait alors d’éviter les nombreux « dérapages » de conception d’outils technologiques, dus à la vision étroite des « experts » qui détiennent une importante formation scientifique ou technique, mais cette dernière étant dépourvue de toute réflexion humaniste (Vicente 2004 : 44–50) :

Les désagréments et les frustrations que nous éprouvons quotidiennement en utilisant nos appareils et nos gadgets sophistiqués ne constituent que la pointe de l’iceberg. Nous devons prendre conscience que ces problèmes appartiennent aussi aux systèmes complexes responsables de notre bien-être et de notre sécurité. (ibid. : 39)

Une révolution « techno-humaine » (ibid. : 56) s’impose donc et la technologie se doit d’être conçue en meilleure adéquation avec l’être humain qui en fait usage et ses besoins. Dans le cas des traducteurs, il s’agira alors de prendre en considération leur utilisation quotidienne des outils informatiques à leur disposition et les limites qu’ils présentent pour leur proposer une interface qui devienne « une extension naturelle des capacités et habiletés du langagier » (Taravella et Villeneuve 2011, paragr. 51).

En cela, la RV et les interfaces multimodales se présentent comme des solutions ergonomiques prometteuses puisqu’elles fonctionnent déjà très bien au sein de divers domaines professionnels et situations de la vie quotidienne. Leur intégration aux applications de traductique actuelles permettrait alors d’en combler les manques et de prendre en compte le facteur humain. La RV contribue déjà au grand retour de la traduction dictée (TD) dans la pratique de la traduction, qui, selon Gouadec (2007 : 363), est en passe de redevenir la norme :

In all probability, the translation industry in the twenty-first century will go oral. Voice recognition and dictation software is probably going to radically change the way translation is perceived and practiced […] Dictated translation will become the norm once again.

Cependant, un usage multimodal et ergonomique de la TD se doit de dépasser les pratiques des années 1960 et 1970, période à laquelle il était commun de traduire à voix haute. C’est pourquoi nous envisageons l’intégration de la TDI, à la fois à la formation et à l’environnement de travail des traducteurs. L’interaction multimodale vise à améliorer la relation entre le langagier et la technologie dans tous les aspects de son travail (préparation, production, révision des traductions) pour accroître son efficacité, sa créativité et son bien-être physique et psychologique. En faisant la promotion d’une adaptation optimale de la RV et des technologies interactives émergentes (ex. : écrans tactiles et appareils mobiles), la TDI propose donc d’apporter aux environnements de traductique cette dimension humaine tant attendue.


Barabé, Donald (2013) « Société, technologie et traduction : perspectives et impacts », The Journal of Specialised Translation 19: 41–61.

Benis, Michael (2002) « Softly spoken or hard of hearing? », Language International 14, no. 3: 26–9.

Bowker, Lynne (2002) Computer-Aided Translation Technology: A Practical Introduction, Ottawa, University of Ottawa Press.

Brousseau, Julie,Caroline Drouin, George Foster, Pierre Isabelle, Roland Kuhn, Yves Normandin et Pierre Plamondon (1995) « French speech recognition in an automatic dictation system for translators: The TransTalk project », Proceedings of Eurospeech’95, URL : http://www.iro.umontreal.ca/~foster/papers/ttalk-eurospeech95.pdf (consulté le 3 mars 2017).

Brown, P. F., S. F. Chen, S. A. Della Pietra, V. J. Della Pietra, A. S. Kehler et R. L. Mercer (1994) « Automatic speech recognition in machine-aided translation », Computer Speech and Language 8, no. 3: 177–87.

Carl, Michael, Barbara Dragsted et Arnt Lykke Jakobsen (2011) « On the Systematicity of Human Translation Processes », Actes de la conférence de Tralogy, Paris, le 3 mars 2011, URL : http://lodel.irevues.inist.fr/tralogy/index.php?id=103 (consulté le 3 mars 2017).

Ciobanu, Dragoș (2014) « Of Dragons and Speech Recognition Wizards and Apprentices », Revista Tradumàtica 12: 524–38, URL : https://ddd.uab.cat/pub/tradumatica/tradumatica_a2014n12/tradumatica_a2014n12p524.pdf (consulté le 3 mars 2017).

Ciobanu, Dragoș (2016) « Automatic Speech Recognition in the Professonal Translation Process », Translation Spaces 5, no.1: 124–44.

Cohen, Philip, Michael Johnston, David McGee, Sharon L. Oviatt, Josh Clow et Ira Smith (1998) « The Efficiency of Multimodal Interaction: A Case Study », Proceedings of the ICSLP’98, URL : http://web.media.mit.edu/~earroyo/thesis-backup/Thesis-prep/multimodal-interfaces/Alredy-printed/the-efficiency-of-multimodal-interaction.pdf (consulté le 3 mars 2017).

Daiute, Colette (1985) Writing and Computers, Reading [MA], Addison-Wesley.

Désilets, Alain, Marta Stojanovic, Jean-François Lapointe, Richard Rose et Aarthi Reddy (2008) « Evaluating Productivity Gains of Hybrid ASR-MT Systems for Translation Dictation », Proceedings of IWSLT 2008, URL : http://www.mt-archive.info/IWSLT-2008-Desilets.pdf (consulté le 3 mars 2017).

Dragsted, Barbara et Inge Gorm Hansen (2009) « Exploring Translation and Interpreting Hybrids. The Case of Sight Translation », Meta 54, no. 3: 588–604, URL : http://www.erudit.org/revue/meta/2009/v54/n3/038317ar.pdf (consulté le 3 mars 2017).

Dragsted, Barbara, Inge Gorm Hansen et Henrik Selsøe Sørensen (2009) « Experts Exposed », Copenhaguen Studies in Language 38: 293–317.

Dymetman, Marc, Julie Brousseau, George Foster, Pierre Isabelle, Yves Normandin et Pierre Plamondon (1994) « Towards an Automatic Dictation System for Translators: the TransTalk Project », Proceedings of the ICSLP’94, URL : http://arxiv.org/abs/cmp-lg/9409012 (consulté le 8 mars 2017).

Garcia-Martinez, Mercedes, Karan Singla, Aniruddha Tammewar, Bartolomé Mesa-Lao, Ankita Thakur, Anusuya M. A., Michael Carl et Srinivas Bangalore (2014) « SEECAT: ASR & Eye-tracking Enabled Computer-Assisted Translation », Proceedings of the 17th Annual Conference of the European Association for Machine Translation: 81–8, URL : http://www.mt-archive.info/10/EAMT-2014-Garcia-Martinez.pdf (consulté le 8 mars 2017).

Girol, David, Ramón Lopez-Cozar, Zoraida Callejas, Gonzalo Espejo et Nieves Ábalos (2012) « On the Development of Adaptive and User-Centered Interactive Multimodal Interfaces », dans Speech, Image and Language Processing for Human Computer Interaction: Multi-Modal Advancements, Uma Shanker Tiwary et Tanveer J. Siddiqui (dir.), Hershey [PA], Information Science Reference: 262–91.

Gouadec, Daniel (2007) Translation as a profession, Amsterdam, John Benjamins.

Jiménez Ivars, Amparo et Amparo Hurtado Albir (2003) « Variedades de traducción a la vista. Definición y clasificación », TRANS: Revista de Traductología 7: 47–57, URL : http://www.trans.uma.es/trans_07.html (consulté le 3 mars 2017).

Jurafsky, Daniel et James H. Martin (2009) Speech and language processing: An introduction to natural language processing, computational linguistics, and speech recognition (2e éd.), Upper Saddle River [NJ], Pearson Prentice Hall.

Karamanis, Nikiforos, Saturnino Luz et Gavin Doherty (2011) « Translation practice in the workplace: A contextual analysis and implications for machine translation », Machine Translation 25, no. 1: 35–52.

Lippmann, Richard P (1997) « Speech recognition by machines and humans », Speech Communication 22, no. 1: 1–15, URL : http://doi.org/10.1016/S0167-6393(97)00021-6 (consulté le 3 mars 2017).

Mees, Inger M., Barbara Dragsted, Inge Gorm Hansen et Arnt Lykke Jakobsen (2013) « Sound effects in translation », Target 25, no. 1: 140–54, URL : http://openurl.ingenta.com/content/xref?genre=article&issn=0924-1884&volume=25&issue=1&spage=140 (consulté le 3 mars 2017).

Mesa-Lao, Bartolomé (2014) « Speech-Enabled Computer-Aided Translation: A Satisfaction Survey with Post-Editor Trainees », Workshop on Humans and Computer-assisted Translation: 99–103, URL : http://www.aclweb.org/anthology/W14-0315 (consulté le 8 mars 2017).

O’Brien, Sharon (2012) « Translation as human-computer interaction », Translation Spaces 1, no. 1: 101–22, URL : http://doi.org/10.1075/ts.1.05obr (consulté le 8 mars 2017).

Office québécois de la langue française (2007) Grand dictionnaire terminologique, URL : http://www.granddictionnaire.com (consulté le 3 mars 2017).

Oviatt, Sharon (2012) « Multimodal Interfaces », dans The Human-Computer Interaction Handbook: Fundamentals, Evolving Technologies and Emerging Applications (3e éd), Julie A. Jacko (dir.), New York, Lawrence Erlbaum Associates: 405–29.

Oviatt, Sharon et Philip R. Cohen (2000) « Multimodal systems that process what comes naturally », Communications of the ACM 43, no. 3: 45–53.

Oviatt, Sharon, Philip Cohen et Michelle Wang (1994) « Toward interface design for human language technology: Modality and structure as determinants of linguistic complexity », Speech Communication 15: 283–300.

Oviatt, Sharon, Rachel Coulston et Rebecca Lunsford (2004) « When do we interact multimodally? Cognitive load and multimodal communication partners », Proceedings of the ICMI’04, New York, ACM Press: 129–36.

Oviatt, Sharon, Rebecca Lunsford et Rachel Coulston (2005) « Individal differences in multimodal integration patterns: What are they and why do they exist? », CHI’05 Proceedings, New York, ACM Press: 241–9.

Oviatt, Sharon et Robert van Gent (1996) « Error resolution during multimodal human-computer interaction », Proceedings of the ICSLP’96: 204–7.

Rabiner, Lawrence (1997) « Applications of speech recognition in the area of telecommunications », Proceedings of the IEEE’97, URL : http://www.ece.ucsb.edu/Faculty/Rabiner/ece259/Reprints/341_telecom applications.pdf (consulté le 3 mars 2017).

Reddy, Aarthi et Richard Rose (2010) « Integration of Statistical Models for Dictation of Document Translations in a Machine Aided Human Translation Task », IEEE Transactions on Audio, Speech and Language Processing 18, no. 8: 1–11, URL : http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=05393062 (consulté le 3 mars 2017).

Romero-Fresco, Pablo (2011) Subtitling Through Speech Recognition: Respeaking, Manchester, St. Jerome.

Schalkwyk, Johan, Doug Beeferman, Bill Byrne, Ciprian Chelba, Mike Cohen, Maryam Garret et Brian Strope (2010) « Google Search by Voice: A Case Study », Advances in Speech Recognition, New York, Springer: 61–90, URL : http://static.googleusercontent.com/external_content/untrusted_dlcp/research.google.com/en/us/pubs/archive/36340.pdf (consulté le 3 mars 2017).

Seaman, Lee (2002) « Voice Recognition for Translators, Or Why I Started Talking To My Computer », JLD Times. Newsletter of the Japanese Language Division of the American Translators Association: 6–8.

Shneiderman, B. (1998). Designing the User Interface: Strategies for Effective Human-Computer Interaction (3rd ed.). Reading, Massachusetts : Addison-Wesley.

Stroman, John (2002) « Translation And Voice Recognition Software », JLD Times. Newsletter of the Japanese Language Division of the American Translators Association: 1–5.

Taravella, Anne-Marie et Alain Villeneuve (2011) « Aspects humains des technologies langagières dans l’organisation », Actes de la conférence de Tralogy, Paris, le 3 mars 2011, URL : http://lodel.irevues.inist.fr/tralogy/index.php?id=134 (consulté le 3 mars 2017).

Taravella, Anne-Marie et Alain Villeneuve (2013) « Acknowledging the needs of computer-assisted translation tools users: the human perspective in human-machine translation », The Journal of Specialised Translation 19: 62–74, URL : http://www.jostrans.org/issue19/art_taravella.pdf (consulté le 3 mars 2017).

Torres-Hostench, Olga, Joss Moorkens, Sharon O’Brien et Joris Vreeke (2017). Testing interaction with a Mobile MT post-editing app. Translation & Interpreting 9, no. 2:138-150, ULR: http://www.trans-int.org/index.php/transint/article/view/645/291 (consulté le 8 aout 2017).

Vicente, Kim (2004) Le facteur humain : réinventer notre rapport à la technologie, trad. N. Guillet, Outremont, Logiques.

Vidal, Enrique, Francisco Casacuberta, Luis Rodríguez, Jorge Civera et Carlos D. Martínez Hinarejos (2006) « Computer-assisted translation using speech recognition », IEEE Transactions on Audio, Speech and Language Processing 14, no. 3, URL : http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=01621206 (consulté le 3 mars 2017).

Williamson, Julie R., Andrew Crossan et Stephen Brewster (2011) « Multimodal Mobile Interactions: Usability Studies in Real World Settings », Proceedings of the 13th International Conference on Multimodal Interaction: 361–8.

Wu, Ming (2008) The Most Cost-effective Way of Translation: Combined Use of Digital Recorders and Dragon Voice Recognition System, Diaporama d’une présentation aux Nations-Unies à Nairobi, le 4 aout. Inédit.

Zapata, Julián (2012) Traduction dictée interactive : intégrer la reconnaissance vocale à l’enseignement et à la pratique de la traduction professionnelle, Mémoire de maîtrise, Université d’Ottawa, URL : http://www.ruor.uottawa.ca/en/bitstream/handle/10393/23227/Zapata Rojas_Julian_2012_these.pdf?sequence=1 (consulté le 3 mars 2017).

Zapata, Julián (2016) Translators in the Loop: Observing and Analyzing the Translator Experience with Multimodal Interfaces for Interactive Translation Dictation Environment Design, Thèse de doctorat, University of Ottawa, URL : https://www.ruor.uottawa.ca/bitstream/10393/34978/1/Zapata_Rojas_Julian_2016_thesis.pdf (consulté le 3 mars 2017).

Zapata, Julián et Andreas S. Kirkedal (2015) « Assessing the Performance of Automatic Speech Recognition Systems When Used by Native and Non-Native Speakers of Three Major Languages in Dictation Workflows », Proceedings of the 20th Nordic Conference of Computational Linguistics: 201–10, URL : http://www.aclweb.org/anthology/W15-1825 (consulté le 3 mars 2017).

Zapata, Julián et Jean Quirion (2016) « La traduction dictée interactive et sa nécessaire intégration à la formation des traducteurs », Babel 62, no. 4: 531–51, URL : https://benjamins.com/#catalog/journals/babel.62.4.01zap/details (consulté le 3 mars 2017).

Zapata, Julián, Sheila Castillo et Joss Moorkens (2017) « Translation Dictation vs. Post-editing with Cloud-based Voice Recognition: A Pilot Experiment », Proceedings of The Sixteenth MT Summit Conference, Nagoya, Japon, 18-22 septembre, 2017.


[1] Dans le présent article, le genre non marqué, c’est-à-dire le masculin, est utilisé uniquement pour alléger le texte et il inclut le genre féminin là où il y a lieu.

[2] L’ergonomie est « l’étude scientifique des conditions (psychophysiologiques et socioéconomiques) de travail, de l’adaptation des outils, postes de travail aux utilisateurs, des relations entre l’utilisateur et la machine » (Petit Robert de la langue française en ligne, 2017).

[3] Nous reconnaissons qu’il existe une terminologie variée pour désigner la notion que nous choisissons d’appeler « reconnaissance vocale ». En français, certains chercheurs s’y référeront à l’aide de termes tels que « reconnaissance automatique de la parole », « reconnaissance de la voix » ou simplement en employant l’acronyme anglais « ASR » (Automatic Speech Recognition). Pour notre part, nous nous appuyons sur la terminologie et la définition proposées par l’Office québécois de la langue française.

[4] Selon une enquête menée par Verástegui en 2009 auprès des participants de la Réunion internationale annuelle sur la traduction et la terminologie assistées par ordinateur (JIAMCATT) regroupant des représentants de grands services de traduction. Les résultats de l’enquête ont été distribués par l’auteur via LISTSERV aux participants de JIAMCATT 2009, mais n’ont fait l’objet d’aucune publication.

[5] Pour en savoir plus sur les autres résultats de l’expérience, nous renvoyons le lecteur à Zapata (2016).

About the author(s)

Julián Zapata is Doctor of Philosophy in Translation Studies, and member of the Association of Part-time Professors of the University of Ottawa. Founder and President of InTr Technologies.

Elizabeth C. Saint is a PhD candidate in Translation Studies at the University of Ottawa. Her research focuses on terminology use and implantation within language planning contexts. Her thesis proposes an online collaborative model between the terminologists and the speakers, final users of the terms (enbonstermes.com).

Email: [please login or register to view author's email address]

©inTRAlinea & Julián Zapata & Elizabeth C. Saint (2017).
"Traduire à voix haute la traduction dictée interactive comme solution ergonomique"
inTRAlinea Volumes
Edited by: {specials_editors_volumes}
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2261

Corpora and Literary Translation

By Titika Dimitroulia & Dionysis Goutsos (Aristotle University of Thessaloniki & National and Kapodistrian University of Athens, Greece)

©inTRAlinea & Titika Dimitroulia & Dionysis Goutsos (2017).
"Corpora and Literary Translation"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2260

The Digital Humanities are attracting an increasing amount of attention in the humanities, the social sciences and the arts; often leading to transnational, collaborative and interdisciplinary projects which are also increasingly based on corpora (Crompton, Lane and Siemens 2017; Bernard et Bohet 2017; Longhi 2017; Klein and Gold 2016; Jockers 2013; Warwick, Terras and Nyhan 2012). Digital Literary Studies, which use a wide range of methodologies, explore the latent hermeneutic potential of related fields, by combining qualitative and quantitative methods which privilege corpora (Earhart 2015; Ganascia 2015; Hoover, Culpeper and O’Halloran 2014; Price and Siemens 2013; Schreibman and Siemens 2008; Marchionne, online). This preference is, in a way, reminiscent of the ‘prehistory’ of corpora, which was rooted in concordances of religious and sacred texts rather than texts with a practical purpose (Gigot 1910; Jones 2016).

The resistance of literary translation to integrating corpora in its practice, analysis and teaching is due both to the relative absence of literary parallel corpora (Olohan 2004) and to the reluctance of researchers in both literary and literary translation studies to investigate the quantitative analysis of literary texts – be they originals or translations (Porsdam 2011). Nevertheless, a corpus-based approach is gaining an increasing amount of ground in literary translation studies (Zubillaga, Sanz and Uribarri 2015; Rybicki 2012; Ji 2012), after being explored in its inception by such researchers as Bernardo (1981), Baker (1996, 1999, 2000), Kenny (2000), Zanettin (2001 and 2000) and Bosseaux (2004), among others. This development is also supported by new tools, like CATMA (http://www.catma.de), the cultural texts annotation tool, TRADUXIO (http://traduxio.hypertopic.org; Goncharova and Lacour 2011), participative platform for cultural texts translators or QU.IT (http://www.quit.unibo.it;  Zotti, 2016), a parallel database useful both to translation practice and teaching. Thus, literary bi-text interfaces and tools can lead the way to complex approaches and the use of corpora in literary translation, while the traditional concordances and parallel corpora continue to be of great help not only to translators, but also to researchers and translation teachers, as this issue highlights.

This is a period when textual material, and particularly literary texts, are flourishing on the internet and can be used for the compilation of parallel corpora (Zanettin 2002). Tools are also increasingly simpler and more effective to use, ranging from OCR to alignment, concordancing, annotating and visualizing, while the new perspectives offered by interdisciplinary projects in Corpus-based Translation Studies make the use of corpora in literary translation practice, research and teaching both a choice and a necessity. Corpus-based and corpus-driven approaches (see below) to literary translation, beyond linguistic and stylistic research, are found in macro-level and/or diachronic analyses, among others, of norms, translation history and terminology management in literary or comparative studies and can reveal the relation between original and translated literary production at many levels. Furthermore, they can contribute to research on literary and cultural transfer by doing justice to its complexity and its multiple relationships to culture and society in a global context.  

The global context is also shaped by the discussion on World Literature or world literatures (Apter 2006, 2013; Damrosch 2006). In this discussion, firstly, fiction is created with regard to the place or even the possibility of translation in it (Apter 2013) and, secondly, the sociological translation perspective applied at a macro-level has to be completed by a close analysis of the texts (Casanova 2004). Parallel and comparable literary corpora, explored from different points of view, and, most importantly, including all genres of literary texts, can contribute to this discussion, dealing with the relation between language, nation, minorities and all kinds of collectivities, their linguistic and cultural universals and particularities. Research on literary translation based on or driven by big corpora is an unexplored domain, which can enrich Translation Studies in their relation to other disciplines (Ganascia, 2015).

This issue aims to highlight the challenges of corpus-based literary translation, as well as its possible itineraries in the future, from a historical, theoretical or practical perspective. In the opening paper Gerard Lynch sets the scene by placing the analysis of literary texts through corpora within the broader context of research involving the computational and statistical analysis of written texts. He starts from a historical overview, by identifying an earlier phase that focused on the so-called translationese or translation’s interlanguage, following Baker’s (1993) pioneering work. The next phase includes, according to Lynch, stylometric studies on literary translation, giving emphasis to micro-features such as those used in author attribution or stylistic profiling, more generally. The current work belongs to the third phase of research, which involves big data and machine learning methods. Lynch discusses relevant work on literary translation in each phase in order to point out important methodological and theoretical considerations, including the need for more widely accessible machine learning platforms, the restrictions imposed by the limited availability of relevant literary material and the necessity to combine macro-level approaches with more fine-grained analyses. In his view, the interaction of literary translation projects with other research in digital humanities has the potential to investigate “trends in stylistic variation that may transcend individual translator’s choices”.

The papers that follow offer four case studies in the analysis of literary translations. Mojca Schlamberger Brezar focuses the discussion on how parallel literary translation corpora can be exploited for a variety of purposes, by showing how corpora like FraSloK, a literary translation parallel corpus involving French and Slovene, are placed within the larger frame of language resources available for the two languages in question. She then moves on to a minute analysis of the frequency of connectives such as the French mais, cependant etc, which have been extensively discussed in the linguistic literature under various names such as discourse, functional or pragmatic markers etc. (see e.g. Fedriani and Sansó 2017), along with their translation variants in the texts of the corpus. Her analysis shows how quantitative considerations can shed light on an author’s poetics and highlights the usefulness of literary corpora, which, because of their nature, include several registers and thus offer a wider spectrum of lexical choices both to native speakers and L2 learners of a language.

In her contribution Adriana Mezeg singles out a specific construction, namely the use of past participles in sentence initial non-finite clauses, in order to identify translation strategies with reference to the same parallel corpus. What is particularly important is the combination of automatic and semi-automatic methods, for example the use of a qualitative functional analysis of its syntactic and semantic properties for the identification of the construction in question and its frequency. The translation process seems to favour, in her findings, explicitation strategies both in syntactic terms and semantic relations, the latter bringing to the fore implicit relations through translation.  The implications are obvious for the linguistic analysis of each language and their contrastive relations, as well as pedagogical and professional applications.

Elaine Ng’s contribution is a close analysis of four Chinese translations of Hemingway’s The Old Man and the Sea. She specifically employs Simpson’s (1993) model of point of view and focuses on the expression of modality through deontic, epistemic and other modals, as found in the original and the translated texts. Her meticulous analysis bears out the translation shifts occurring in the rendering of a literary text in Chinese and allows the identification of translation strategies, such as omission or modification that may be crucial for the overall patterning of speech and thought presentation in the translated texts. Although, as pointed out, extra-textual information is needed in order to better understand the translators’ choices, quantitative corpus analyses such as this can reveal recurring patterns of translation activity and stylistic choice. Despite its small sample, this is a good example of how close textual analysis can draw from quantitative, and more specifically, corpus-based methods in order to reveal what takes place in the translation process, both on the level of individual translators and that of a particular language.

Finally, Rudy Loock shifts the attention to a rather neglected type of parallel corpora, namely a learner corpus which involves translation tasks, in his case from English to French performed by advanced students of English, having French as their native language. He thus brings the discussion back to the question of translation’s interlanguage or the first phase of relevant research in Lynch’s taxonomy, by carefully comparing student texts with English and French corpora as regards two specific linguistic features, namely derived adverbs in -ly vs. –ment and existential there vs. il y a constructions. The findings are also correlated with the assessment of translation quality by independent evaluators. This exploratory and thus highly original type of research is valuable not so much for its definitive results, but rather for the methodological and theoretical issues it raises, including the question of the correlation between intra-language differences as a whole and translation quality.

This special issue on the use of corpora in the study of literary translation concludes with a paper by Federico Zanettin, which aptly wraps up the issues involved. In particular, Zanettin places the corpus analysis of literary translation within the tradition of computer-assisted literary studies, pointing out the difference between qualitatively- and quantitatively-driven or stylometric approaches, in a way reminiscent of Tognini-Bonelli’s (2001) famous distinction between corpus-based and corpus-driven analysis. His extensive overview of corpus studies that focus on literary texts indicates the spectrum of methods employed, the range of linguistic features examined and the diverse findings produced. His paper draws the implications of these studies for exploring literary translators’ style by means of corpora and the question of translation universals (see Baker 1993), as well as translation criticism.


Baker, Mona (1993) “Corpus Linguistics and Translation Studies — Implications and Applications” in Text and Technology: In Honour of John Sinclair, Mona Baker, Gill Francis and Elena Tognini-Bonelli (eds), Amsterdam/Philadelphia, Benjamins, 233-50.

----, (1996) “Corpus-based Translation Studies: The Challenges that Lie Ahead” in Terminology, LSP and Translation: Studies in Language Engineering in Honour of Juan C. Sager, Harold Somers (ed.),. Amsterdam/Philadelphia: Benjamins, 567-86.

----, (1999) “The Role of Corpora in Investigating the Linguistic Behaviour of Professional Translators”, International Journal of Corpus Linguistics, 4(2): 281-98.

----, (2000) “Towards a Methodology for Investigating the Style of a Literary Translator”, Target 12 (2), 241–66.         

Bernard, Michel et Bohet, Baptiste (2017) Littérométrie. Outils numériques pour l’analyse des textes littéraires, Paris, Presses Sorbonne Nouvelle.

Bernardo, Aldo (1981) “Maximizing Computer Assistance in Literary Translation: Petrarch's Familiares” in Translation Spectrum: Essays in Theory and Practice, Marilyn Gaddis Rose (ed.), New York, State University of New York Press, 74-80.

Bosseaux, Charlotte (2004) “Point of View in Translation: A Corpus-based Study of French Translations of Virginia Woolf’s To the Lighthouse”, Across Languages and Cultures, 5(1): 107–22.

Crompton, Constance, Lane, Richard J., Siemens, Ray (2017) Doing Digital Humanities. Practice, Training, Research, London/New York, Routledge.

Earhart, Amy E.  (2015) Traces of the Old, Uses of the New The Emergence of Digital Literary Studies. Michigan, University of Michigan Press.

Fedriani, Chiara & Sansó, Andrea (eds) (2017) Pragmatics Markers, Discourse Markers and Modal Particles: New Perspectives, Amsterdam/Philadelphia, Benjamins.

Ganascia Jean-Gabriel. (2015) “The Logic of the Big Data Turn in Digital Literary Studies”, Frontiers / Digit. Humanities, 2 (7), online. DOI: 10.3389/fdigh.2015.00007

Gigot, Francis E. (1910) “Hugh of St. Cher” Catholic Encyclopedia. New York, Robert Appleton.

Goncharova, Yuliya and Lacour, Philippe (2011) “TraduXio : Nouvelle expérience en traduction littéraire”, Traduire online, 225 | 2011. DOI : 10.4000/traduire.94

Ji, Meng (2012) “Hypothesis Testing in Corpus-based Literary Translation Studies”. In Quantitative Methods in Corpus-based Translation Studies, Michael Oakes and Meng Ji (eds), Amsterdam/Philadelphia, Benjamins, 53–72.

Jockers, Matthew L. (2013) Macroanalysis: Digital Methods and Literary History, Urbana-Champaign, IL, University of Illinois Press.

Jones, Steven E. (2016) Roberto Busa, S. J., and the Emergence of Humanities Computing: The Priest and the Punched Cards, London/New York, Routledge.

Kenny, Dorothy (2000) Lexis and Creativity in Translation: A Corpus-based Study. Manchester, St. Jerome Publishing.

Klein, Lauren F. and Gold, Matthew K. (2016) “Digital Humanities: The Expanded Field”, Debates in the Digital Humanities, http://dhdebates.gc.cuny.edu/debates/2

Longhi, Julien (2017) “Humanités, numérique : des corpus au sens, du sens aux corpus‪” Questions de communication 31 (1) : 7-17. https://www.cairn.info/revue-questions-de-communication-2017-1-page-7.htm.

Marchionne, Ilaria (online) “I quattro livelli della rivoluzione digitale in ambito letterario. Capitolo uno: i corpora linguistici”, Digital Writing Lab, Gruppo di ricerca del Communication Strategies Lab, http://www.csl.unifi.it/dwl/riflessi/i-quattro-livelli-della-rivoluzione-digitale-in-ambito-letterario-capitolo-uno-i-corpora-linguistici/

Olohan, Maeve (2004) Introducing Corpora to Translation Studies, London/New York,  Routledge.

Porsdam, Helle (2011) “Too much ‘digital’, too little ‘humanities’? An attempt to explain why many humanities scholars are reluctant converts to digital humanities”. In Arcadia Project, Cambridge, MA, Cambridge University Laboratory, University of Cambridge, http://arcadiaproject.lib.cam.ac.uk/docs/DigitalHumanities.pdf

Price, Kenneth M. and Siemens, Ray (2013) Literary Studies in the Digital Age. An Evolving Anthology, New York, NY, MLA Commons, Modern Language Association, https://dlsanthology.mla.hcommons.org/

Rybicki, Jan (2012) “The Great Mystery of the (almost) Invisible Translator. Stylometry in Translation” in Quantitative Methods in Corpus-Based Translation Studies, Michael Oakes and Meng Ji (eds), Amsterdam/Philadelphia: Benjamins, 231-48.

Schreibman, Susan & Siemens, Ray (eds) (2008) A Companion to Digital Literary Studies. Oxford, Blackwell, http://www.digitalhumanities.org/companionDLS/


---, and Unsworth, John (eds) (2004) A Companion to Digital Humanities, Oxford, Blackwell, http://www.digitalhumanities.org/companion

Simpson, Paul (1993) Language, Ideology and Point of View, London/New York, Routledge.

Tognini-Bonelli, Elena (2001) Corpus Linguistics at Work,  Amsterdam/Philadelphia, Benjamins.

Warwick, Claire, Terras, Melissa and Nyhan, Julianne (eds) (2012) Digital Humanities in Practice, London, Facet.

Zanettin, Federico (2000) “Parallel corpora in translation studies. Issues in corpus design and analysis” in Intercultural Faultlines, Maeve Olohan (ed.), Manchester, St Jerome, 105–18.

Zanettin, Federico (2001) IperGrimus, inTRAlinea Monographs, http://www.intralinea.org/monographs/zanettin


Zotti, Valeria (2016) “QU.IT une plateforme électronique d’aide au travail des traducteurs littéraires. La Francesistica italiana à l'ère du numérique”, Publifarum, 25, http://publifarum.farum.it/ezine_articles.php?id=337.

Zubillaga, Naroa, Sanz, Zuriñe  and Uribarri, Ibon (2015) “Building a trilingual parallel corpus to analyse literary translations from German into Basque” in New Directions in Corpus-based Translation Studies, Claudio Fantinuoli and Federico Zanettin (eds.), Berlin, Language Science Press, 71–93.

About the author(s)

Titika Dimitroulia is Associate Professor of Translation Studies in the School of French at Aristotle University of Thessaloniki and Director of the Greek National School of Public Administration and Local Government. She coordinates AUF (Agence Universitaire de la Francophonie) and is a member of the Hellenic Terminology Network (https://ec.europa.eu/greece/el-diktyo_el) and the CLARIN infrastructure (European Research Infrastructure for Language Resources and Technology, http://www.clarin.gr). She is also Director of the Digital Humanities Laboratory (http://www.digitalhumanitieslab.eu) and responsible for the translation curriculum and tutor at the Training Programme for Greek-speaking translators of the Academy of Athens. She has received the EKEMEL (European Translation Centre-Literature and Human Sciences) translation award in 2008 and has collaborated with several Greek newspapers as well print and electronic journals as a literary critic. She has published widely on literature, translation and digital literary studies, as well the following books, among else: Literary Translation. Theory and Practice (2015, in Greek); Digital Literary Studies (2015, in Greek).

Dionysis Goutsos is Professor of Text Linguistics at the National and Kapodistrian University of Athens. He has also taught at the University of Birmingham (UK) and the University of Cyprus. He has written several articles on text linguistics and discourse analysis, translation studies and corpus linguistics, as well as the following books, among else: Modeling Discourse Topic (1997), Discourse Analysis: An Introduction (1997/2004), The Discourse of Translation (2001, in Greek) and Language: Text, Variety, System (2012, in Greek). He has been research co-ordinator for the research projects leading to the compilation of the Corpus of Greek Texts (http://www.sek.edu) and the Diachronic Corpus of Greek of the 20th Century (http://greekcorpus20.phil.uoa.gr/).

Email: [please login or register to view author's email address]

©inTRAlinea & Titika Dimitroulia & Dionysis Goutsos (2017).
"Corpora and Literary Translation"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2260

Strange bedfellows

Shifting paradigms in the corpus-based analyses of literary translations

By Gerard Lynch (Independent Scholar, Ireland)

Abstract & Keywords

Although the application of computational and statistical approaches in the analyses of literary translations has gained considerable traction of late, there is still a good deal of progress to be made on bridging the divide between the cross-disciplinary nature of such analyses and the traditions of translation studies and the computing sciences. This article attempts to document issues and methodological divergences which arise in such work, collaborative or otherwise, drawing upon studies across the digital humanities, stylometry, corpus linguistics and translation studies, and identifying areas in which further progress could be achieved based on a mutual collaborative spirit between disciplines.

Keywords: machine learning, stylometry, corpu-based studies

©inTRAlinea & Gerard Lynch (2017).
"Strange bedfellows Shifting paradigms in the corpus-based analyses of literary translations"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2257

1. Introduction

The application of computational and/or statistical analyses to the written text is a practice that dates back to as early as the 19th century. In 1887 Mendenhall's pioneering word-length distribution analysis of British authors Dickens and Thackeray was published in Science, followed by an expanded analysis in which he compared Shakespeare with his contemporaries (Mendenhall 1901). This experiment was inspired by earlier personal correspondence from British logician (Augustus) de Morgan, who touched on the idea of using word length measures to determine the author of the Pauline epistles.[1]

From these early steps towards solving a number of age-old ‘who wrote it?’ questions grew a field which now spans disciplinary boundaries. Zipf (1935) defined the eponymous law, one of the earliest forays into corpus linguistics, stating that the frequency of any word in a natural language corpus is inversely proportional to its rank; a phenomenon that had also been noted in relation to population growth in cities (Auerbach 1913).

British statistician George Udny Yule examined sentence length distributions in the works of Coleridge and Bacon (Yule 1944) and carried out an attribution exercise on the works of Thomas à Kempis, spearheading a second wave of authorship attribution exercises using statistical methods. After the Second World War, a number of scholars ushered in a new age of computational language analysis and one should not underestimate the value of Father Roberto Busa’s work. Considered the first digital humanist, his collaboration with Thomas Watson from IBM on a digital concordance of texts by Thomas Aquinas was ground-breaking.

The next milestone in language and computation came through the work of Warren Weaver on a system facilitating automated translation of Russian into English, a Cold War-era US intelligence project that would create the field of machine translation. In its turn, this would generate much of the research methodology and fund the resources, firmly establishing computational linguistics and natural language processing as a discipline in the twentieth century.[2]

Complementing this pioneering work came a study which diverted the attention back to authorship attribution, namely the work of Mosteller and Wallace (1963) on the authorship attribution of the Federalist papers. This was the first such work to use a computer to analyse texts, previous works having relied on statistical methods which were computed by hand.

In today’s era of Google Translate, culturomics and big data,[3] computational analyses of literary translation are still a relatively niche endeavour. Although the application of corpus linguistic methods has become commonplace in translation studies since the 1990’s,[4] there appears to be a smaller body of work concerned with applying state-of-the-art techniques in natural language processing to answering questions about literary translations in particular. This article outlines a number of factors which may be hampering research collaboration in this area and makes a number of suggestions for possible future collaborative studies.

2. Trends in translation stylometry

This section attempts to contrast the methodological treatment of corpus studies in literary stylometry from the perspective of translation studies and corpus linguistics, the digital humanities and literary stylometry, and the new emergent paradigm which employs text classification and machine learning approaches for research purposes.

2.1 First phase: Of translation universals and early corpus studies

Since Gellerstam (1986) coined the term translationese as a descriptor for what Frawley (1984) referred to as the third code,[5] there has been interest in the application of corpus linguistic methods to questions of translation stylometry, in particular quantifying distinctions between translationese and non-translated language. Baker (1993, 2000) defined a framework for investigating the stylistic properties of parallel translations, giving examples from the literary domain. While not excessively sophisticated, the computational methodology used, based on the analysis of word frequencies, was not common at the time in the translation studies community. Baker posed three main questions when investigating stylistic variation in literary translation: 1) whether a translator’s preference for particular linguistic constructions is independent of the style of the original author, 2) whether the stylistic choices are independent of the general preferences of the target language, and if so, 3) whether they could be explained with reference to the social, cultural or ideological position of the translator.

Adopting a similar methodology, Mikhailov and Villikka (2001) focused on Finnish translations from Russian using statistical measures from corpus linguistics, reporting consistencies in translator’s style across separate translations. Their approach results in the conclusion that modals and sentence length define translator’s style, and they do not consider the translator’s social background or other reasons for the stylistic variation. Olohan (2001) focuses on the explicitation translation universal in her work which focused on optional items in translated English, comparing a corpus of translations with the British National Corpus using statistical tests. Other substantial corpus-based studies include Kenny (2001) on lexical choice in translation, Laviosa-Braithwaite (1997) on simplification (see Zanettin (2013) for an excellent overview).

More recently, Saldanha (2011) expands upon Baker’s blueprint for corpus linguistic research in translation studies, investigating a corpus of translations from Portuguese and Spanish into English, with a focus on stylistic features such as italicized text, preservation of cultural expressions from the source language and use of the that connective. She also compares the individual frequencies of each feature per translator with a bilingual reference corpus in order to situate the feature variation in a larger linguistic context. Winters (2007) work on German translations of F. Scott Fitzgerald’s The Beautiful and the Damned identified speech-act reporting verbs as markers of translator’s style. Li, Zhang, and Liu (2011) investigated translator style from a more sociological perspective, contrasting the translation of the Chinese epic Hongloumeng by a British sinologist with that of an official Chinese government-sanctioned translator, focusing on type-token ratio and related corpus measures to profile stylistic divergence and thereby translation ideology.

As evidenced by the literature, these studies tend to be fine-grained accounts of the translation of one particular work or of a pair of translators, paying close attention to any cultural or ideological bias or effects in the translation which can be attributed to the translator. This can contrast somewhat with the studies in literary stylometry in the next section, which can be indeed fine-grained but tend not to focus on cultural and ideological markers in translation, but rather on issues of stylistic development over time, influence and identifying divergent patterns in joint translations.

2.2 Second Phase: Burrowing into Translation and Corpora++

Alongside the research in the translation studies literature, scholars in the areas of digital humanities and literary stylometry have been applying corpus linguistic methods to questions of stylometry for quite some time. Although translations were not the main focus of the discipline, there are a number of scholars who focused on tasks in this space. Commonalities in these studies include the usage of non-standard statistical metrics such as John Burrow’s Delta method[6] and a focus on the stylometric distinction of character idiolects in an author’s canon, which is often extended to translation. Another aspect of studies in this field which refer to translation is the focus on pre-20th century texts, presumably for reasons of data accessibility. Following on from his pioneering study on the analysis of character idiolects in the works of Jane Austen (Burrows 1987), in later work Burrows (2002b) inadvertently established a breakaway subfield of literary translation stylometry by applying his Delta metric to twenty translations from Latin to English of Roman poet Juvenal, identifying a seventeenth century translation by Thomas D’Urfey as the most similar to all of the other translations. Burrows concluded that this translation may have been used as a reference translation for later works. The application of stylometry to non-prose text is generally less common, although Pantopoulos (2009) investigated the stylistic characteristics of four English-language translators of modern Greek poetry using corpus linguistic methods, and recent work by Herbelot (2014) investigated a distributional semantics approach to modelling poetry.

Rybicki (2006) took up the mantle of literary translation stylometry and applied Burrow’s Delta metric to character idiolects in translations of the Polish epics of Sienkiewikz. He found similarly distinct clustering of character idiolects[7] in translations and in original texts. Lynch and Vogel (2009) applied Rybicki’s methodology coupled with the χmetric to translations of Henrik Ibsen in English and German and their source texts, and observed similar patterns of distinctive characterization in source texts and translations alike. Rybicki (2012) continued his work on translation stylometry with Burrow’s Delta, finding that texts tended to cluster by author rather than translator in a corpus of English to Polish literary translations using frequent words as features, a result which he claims verifies Venuti (1995) and his theory of a translator’s invisibility. Later work by Rybicki and Heydel (2013) investigated a collaborative translation between two translators working on the same translation of Woolf’s Night and Day into Polish, where one translator completed the work of the other after her death, successfully identifying the point where the new translator took up the task. Related work on Slavic languages was carried out by Grabowski (2011, 2013) who investigated translation patterning in English, Russian and Polish versions of Nabokov’s Lolita, focusing on sentence length and word type/token distributions in source and target texts.

The majority of stylometric studies on literary translation use relatively shallow textual features. However, a handful of studies go beyond word n-grams and Burrow’s Delta in their scope. Lucic and Blake (2011) use parses from the Stanford Lexical Parser (Chen and Manning 2014) of two English translations of Rainer Maria Rilke and observe differing patterns of negation and adverbial modifiers between the two translations. A novel approach is employed by El-Fiqi, Petraki, and Abbass (2011) who use a network-theoretic linked representation of a text to identify patterns in two translations of the Holy Q’Uran into English, Hung et al. (2010) who use variable-length grams to attribute Chinese translations of 4th century Buddhist texts written in Sanskrit, and Popescu (2011) who investigates translations in a corpus of literary text from Project Gutenberg using string kernels (i.e. sequences of characters) as distinguishing features, but advocates caution in this approach from an interpretability point of view.

The literary stylometric studies which focus on translation have some elements in common with the studies from the translation studies literature, such as the focus on individual translators and authors, although the focal points of the research projects can often differ, with more of an emphasis on attribution, stylistic profiling and language evolution. The next phase of studies share properties including macro-level analysis and discovery of over-arching trends and universals when applied to translated text.

2.3 Third Phase : The Machine (Learning) Age

The adoption of machine-learning approaches to all possible hypotheses in the current era of big data does not exclude questions of corpora and translation and this phase follows the pioneering work of Baroni and Bernardini (2006) on a comparable corpus of Italian journalistic text. These methods can identify stylistic patterns in textual corpora with considerable ease and may be useful to identify hitherto unconsidered textual variation. These studies often employ a wide range of textual feature representations, including part-of-speech tokens, corpus statistics such as type-token ratio and lexical richness measures and word n-grams. Commonalities in studies employing such methodology include investigation of the stylistic patterns of individual translators and macro-analyses, often eschewing literary genres for journalistic text and focusing on translations as target language corpora only.

Questions regarding gender effects in translation stylometry have been raised in Shlesinger et al. (2010), following on earlier work by Koppel et al. (2002) which investigated gender differences in literary language using machine learning methodology, and Leonardi (2007), who investigated differences in translator ideology and language by gender. In the experiments on translated text, their classifiers failed to distinguish between male and female translators using frequent word features, with classification results barely above the chance baseline of 50%, although earlier corpus statistics had identified statistically significant features in the overall subcorpora. The corpus used in these experiments consisted of 213 literary extracts translated from more than 30 source languages.[8] From this study, the authors surmise that the application of supervised learning approaches to questions of textual stylometry provide a more rigorous methodology for establishing the efficacy of stylistic features. Conventional statistics identified a set of features which occurred statistically more often in translations from translators of each gender. However, these features ultimately did not prove effective at discriminating the gender of the translator of a text using machine learning methods. The subject of language, gender and translation is one which deserves further investigation, perhaps using a source-language restricted corpus to restrict variability and improve the efficacy of any machine learning experiments by restricting the number of possible factors of influence on textual style.

Lynch and Vogel (2012) applied support vector machine classifiers to a corpus of literary translations in order to detect the source language of a textual segment, an approach which had been considered previously for non-literary text such as Europarl and journalistic articles from The New York Times and Haaretz, among others (van Halteren 2008, Koppel and Ordan 2011). This approach succeeded in distinguishing the source language of a text and identified a number of corpus features and word and POS n-grams[9] which were distinctive between translations from different source languages. However, the study did not rigorously examine whether features were truly source language markers or somehow artefacts of the texts chosen. Recently, this work was replicated (Klaussner et al. 2014) using a completely disjoint set of contemporaneous texts. Comparable results were reported, along with additional distinctive features such as trigrams of part-of-speech tags which had not been employed in the previous study, which allays any fears about the original results being dependent on the particular corpus used.

More recent research using advanced machine learning methods have challenged Venuti’s theory of translator’s invisibility. Forsyth and Lam (2013) investigated parallel translations into English of the correspondence between Vincent Van Gogh and his brother Theo and identified separate stylistic traces of original authorship and translator’s style using both machine learning and corpus linguistic approaches, finding that the stylistic divergence between translations (translation distinctiveness) the was less strong than stylistic divergence between originals (authorial discriminability). This correlates with Vajn (2009), who developed a theory of two-dimensional style as applied to parallel English translations of Plato’s Republic, incorporating authorial and translatorial stylistic features into the descriptive process. Bogdanova and Lazaridou (2014) developed a related study investigating the preservation of authorial style across translation from an authorship attribution perspective and attempted to cluster works by their original author both in the original language and translation, one approach being to translate the translations back to the original language using machine translation and applying stylistic clustering techniques to the machine translation and the original text. Lynch (2014) focuses on the translations of Ivan Turgenev, Anton Chekhov and Fyodor Dostoyevsky by the Victorian-era British translator Constance Garnett and uses support vector machine classifiers to guess the original author of a translation segment by Garnett obtaining high classification accuracy between the source authors, using document statistics such as average sentence length and lexical richness on the textual segments and even better accuracy employing linguistic features such as reporting verbs and part-of-speech clusters. This work was motivated by an article by Remnick (2005) on the translation of Russian literature in which Vladimir Nabokov and other contemporaries were quoted as mentioning that the translations of Garnett had no distinct voice, her translations of Turgenev being indistinguishable from her renderings of Dostoyevsky.

3. Methodological and motivational considerations

3.1 Supervised learning and corpus tools

The state-of-the-art in text analytics research and analysis tends towards supervised learning approaches such as support vector machines (Joachims 1998), Naive Bayes classifiers and neural networks. These have an extra degree of complexity than previous statistical metrics employed in corpus linguistics. However, they can be very useful for discovering unknown patterns in data, when compared with traditional statistical measures such as t-tests and chi-square tests which require prior domain knowledge about which features to measure. Although there are both open-source and proprietary software packages available to carry out these analyses such as R (http://www.r-project.org), WEKA (Frank et a. 2005), and RapidMiner (http://rapidminer.com), these tend to come with a learning curve which may be too steep for curious humanists and corpus linguists who have firmly adopted software packages such as the IMS Open Corpus Workbench (Christ 1994) and the WordSmith tools package (Scott 1996) for use in their descriptive analyses.

There is a growing need for more straightforward machine learning platforms which can be used by a greater subsection of the population at large, and Google’s Prediction API (https://developers.google.com/prediction) is a step in this direction, facilitating the application of machine learning for non-experts, although it does not focus specifically on text processing. The emergence of graphical tools which allow for the creation of text-processing pipelines through a drag-and-drop interface is also a development of note in this space which can aid research in advanced textual stylometry in general. Perhaps a hybrid system combining a collocation-viewer with a graphically controlled machine learning module with language analytics functionality via the Stanford NLP Toolkit (Manning et al. 2014) would be the ultimate general-purpose tool of choice for future studies in translation stylometry, although open-source solutions such as AntConc (Anthony et al. 2013) are approaching this level of functionality.

3.2 Corpora and data management

The field of statistical machine translation applies statistical methods to large parallel translation corpora such as Europarl and the Canadian Hansard corpus in order to derive models to translate from one language to another. Machine translation has been a mainstay of natural language processing research since Weaver’s pioneering paper in the 1940’s and attracts a considerable amount of research funding year-on-year. As a result, a good deal of computational stylistic analysis of translations focuses on corpora such as these, with scant regard for literary translations.

Thus, large scale parallel corpora of literary translations are not as prevalent, with the translation studies literature focusing on small-scale analyses of one work or at most, the complete works of a particular author. As a result, studies which focus on a fine-grained analysis of the work of one author using computational methods can often be subject to critique in the computational linguistics literature, as the tendency is to use as large a corpus as possible to obtain statistically significant and generalisable experimental results.

Another aspect hampering fully parallel analyses of literary translations is the availability of parallel reference translations in online sources such as Project Gutenberg (http:// www.wikisource.org) and WikiSource (http://www.wikiproject.org), which try to improve matters by imposing a wiki-format browsing structure on copyright-expired source texts. Given the effort that can be involved in digitising a print work, it is perhaps unsurprising that, copyright restrictions notwithstanding, it is unusual to find two contemporary translations of the same work on a site such as Project Gutenberg, unless that work is indeed of high literary regard. Due to this accessibility bottleneck, curious computational humanists can quickly become discouraged when searching for corpora for their investigations.

Cheesman et al. (2010) propose the notion of translation arrays, encompassing linked databases of multiple translations of the same source text in various languages, arguing that such data would provide a rich resource for cross-cultural studies in language evolution, in the monolingual case,[10] and general cultural questions also.

Their own related investigations have focused on the visualisation of textual re-use in translations of common texts such as the Bible, the same text examined by Covington et al. (2014) who cluster Biblical translations using dendogram clustering and document-level metrics such as sentence length, average type-token ratio and “idea density” (i.e. the ratio of factual propositions to words). This approach can be extended to any texts for which multiple parallel translations are available.

3.3 Asking questions

Based on the literature surveyed here, the scholar’s own background can tend to influence how they approach a study in translation stylometry, based on the individual norms of their host discipline.

Applying best practices for computational studies can bias the approach of a research project in literary translation stylometry. Such studies tend to adopt a macro-level view into translation stylometry, such as defining characteristics of translationese in a large corpus of reportage (Baroni and Bernardini 2006, Koppel and Ordan 2011), quantifying the existence of translation universals in technical and medical translations (Ilisei and Inkpen 2011) and separating a corpus of technical translations by translation direction ( Kurokawa et al. 2009).

Research stemming from literary studies and translation studies tend to take a more fine-grained approach, usually attempting to ground theories about the cultural background or processes of the translator through an examination of their translation result, often compared with the source, following work by such scholars as Baker and Saldanha. Thus, the application of machine learning and advanced stylometric analysis can often be hampered by the relatively small size of the corpus under examination or the particularly fine-grained nature of the research question.

On the other hand, those in the translation studies field often have first-hand unfettered access to contemporaneous copyrighted translations which are not available to the general public such as the Translational English Corpus (Olohan 2002), which is a considerable advantage in carrying out corpus studies. Thus, cross-discipline collaborations often produce the most interesting experimental design and studies. Examples include projects by computational linguists Marco Baroni and translation studies scholar Silvia Bernardini and collaborations between translator and English professor Jan Rybicki with corpus linguistics scholar Maciej Eder, which can draw on strengths from individual fields and also respect the validity of investigatory norms of the disciplines.

4. Towards a shared future

With this spirit in mind, there are many areas in which both researchers of a quantitative nature and translation studies scholars may collaborate which maximise the skills and talents of each discipline. Within the computational linguistics community, there is a growing interest in the analysis of literary text, including translation, with the establishment of the Workshop on Computational Linguistics for Literature, currently in its third iteration, and issues of collaboration with the humanities have been documented by Hammond et al. (2013). With the relentless drive towards data-driven methods in the digital humanities, it is important not to lose sight of the original tenets of translation studies mainstays, such as Baker, who espouse that corpus-based studies should seek to ground their analyses in the cultural background, bias or preferences of a translator or movement. At the same time, progress in automatic text analytics has resulted in a set of tools and processes which enable detailed, semi-automated linguistic and stylistic analyses of large textual corpora in near real-time.

Neuman et al. (2013) describe computational methods toward the analysis of the quantity of metaphor in text, such approaches may also be applied to literary translations in order to quantify the level of metaphor within these and also apply quantitative approaches in concert with theories of metaphor in translation such as the work by Steiner (2002) Also, any application of Latent Dirichlet Allocation or topic modelling approaches, (Blei et al. 2003)) to translated text may serve to illustrate over-arching trends of interest within parallel translations and may prove a fruitful area of interest for the future.

Other topics which have not received substantial attention to date are questions related to how a translator’s style relates to their own authorial style. There is of course a rather limited pool of translators who are also published authors in their own right. However, Wang and Li (2012) focus on this issue in their study of two Chinese translations of Joyce’s Ulysses. They separate translation effects into those from the source language, such as the post-positioning of adverbial phrases in the target text, from individual lexical choice by an individual translator, manifested in the choice of a dialectically sensitive translation of the verb to know in English and one translator’s systematic overuse of the Chinese verb duo (to stroll or to saunter) in both his translation and his original writing.

Furthermore, the application of syntactic and semantic parsing to literary translations is another area which has received little attention, save previously mentioned work by Lucic and Blake (2011). One possible reason for this gap may be misconceptions about the applicability of software trained on modern textual corpora such as the Wall Street Journal to literary texts,[11] although some basic investigations can confirm resources are available for textual parsing in Latin, (Bamman and Crane 2006), medieval Portuguese, (Rocio et al. 2003) and medieval French (Mazziotta 2010, Stein and Prevost 2013). In a similar vein, van Dalen-Oskam (2012) uses open-source general purpose named-entity recognition tools in her study of the translation of proper names in Dutch novels and their English translations, finding the accuracy of the NER software to be more than sufficient for the needs of the study, although mentions that some customisation of the software would be a welcome addition in order to recognise subtypes of proper names.

Many collaborative endeavours between humanists and computational scientists are commonly built around the idea of creating shared resources as evidenced by the Text Encoding Initiative and related projects (Hockey and Walker 1993). Although there is a rise in the number of data-driven humanities projects and a number of key scholars in the area span both fields quite comfortably, including Jockers (2013) who has investigated large-scale topic models on thematic trends in literature, and the team behind the Google Ngrams project (Michel et al. 2011). A possible study in this space could apply big data analysis methodology to a large corpus of parallel translation with a constant L1 and L2 to investigate trends in stylistic variation that may transcend individual translator’s choices.

On a final note, it may be of interest to investigate the latest machine learning approach du jour, so-called deep learning (Le Cun et al. 2015), which leverages artificial neural networks in multiple layers to learn human-like representations of categories in data. However, these approaches typically expect gargantuan labelled datasets for training and thus may be a little overkill for current purposes.


Altintas, K., Can, F., and Patton, J. (2007). Language change quantification using time-separated parallel transla- tions. Literary and Linguistic Computing, 22(4), 375–393.

Anthony, L., Crosthwaite, P., Kim, T., Marchand, T., Yoon, S., Cho, S.-Y., Oh, E., Ryu, N.-Y., Hong, S.-H., Lee, H.-K., et al. (2013). A critical look at software tools in corpus linguistics. Linguistic Research, 30(2), 141–161.

Auerbach, F. (1913). Das gesetz der bevoelkerungskonzentration. Petermanns Geographische Mitteilungen, Vol. LIX 

Baker, M. (1993) Corpus Linguistics and Translation Studies. Implications and Applications, in Baker, M., J. Francis and E. Tognini Bonelli, Text and technology: in honour of John Sinclair, Amsterdam and Philadelphia: John Benjamins, 233-250.

Baker, M. (2000) ‘Towards a methodology for investigating the style of a literary translator’, Target, 12(2), 241-266.

Bamman, D., and Crane, G. (2006). The design and use of a Latin dependency treebank. In Proceedings of the Fifth Workshop on Treebanks and Linguistic Theories (TLT2006), pp. 67–78.

Baroni, M., and Bernardini, S. (2006). A new approach to the study of translationese: Machine-learning the difference between original and translated text. Literary and Linguistic Computing, 21(3), 259.

Blei, D. M., Ng, A. Y., and Jordan, M. I. (2003). Latent dirichlet allocation. Journal of Machine Learning Research, 3, 993–1022.

Bogdanova, D., and Lazaridou, A. (2014). Cross-Language Authorship Attribution. In Language Resources and Evaluation, Reykavik, Iceland

Burrows, J. (2002a). ’Delta’: a measure of stylistic difference and a guide to likely authorship. Literary and Linguistic Computing, 17(3), 267.

Burrows, J. (2002b). The Englishing of Juvenal: computational stylistics and translated texts. Style, 36(4), 677-699.

Burrows, J. F. (1987). Computation into criticism: A study of Jane Austen’s novels and an experiment in method. Clarendon Press Oxford.

Candel-Mora, M. A., and Vargas-Sierra, C. (2013). An analysis of research production in corpus linguistics applied to translation. Procedia-Social and Behavioral Sciences, 95, 317–324.

Cheesman, T., Thiel, S., Flanagan, K., Geng, Z., Ehrmann, A., Laramee, R. S., Hope, J., and Berry, D. M. (2010). Translation Arrays: Exploring Cultural Heritage Texts Across Languages. In Proceedings of Digital Humanities 2012, Kings College London.

Chen, D. and C. D Manning (2014) A Fast and Accurate Dependency Parser using Neural Networks. Proceedings of EMNLP 2014.

Christ, O. (1994). The IMS corpus workbench technical manual. Institut fuer Maschinelle Sprachverarbeitung, Universitaet Stuttgart.

Covington, M. A., Potter, I., and Snodgrass, T. (2014). Stylometric classification of different translations of the same text into the same language. Literary and Linguistic Computing. Advance Access

El-Fiqi, H., Petraki, E., and Abbass, H. (2011). A computational linguistic approach for the identification of translator stylometry using Arabic-English text. In Fuzzy Systems (FUZZ), 2011 IEEE International Conference on, pp. 2039-2045. IEEE.

Forsyth, R. S., and Lam, P. W. Y. (2013). Found in translation: To what extent is authorial discriminability preserved by translators?. Literary and Linguistic Computing.29(2),199-217

Frank, E., Hall, M., Holmes, G., Kirkby, R., Pfahringer, B., and Witten, I. (2005). Weka: A machine learning workbench for data mining. Data Mining and Knowledge Discovery Handbook: A Complete Guide for Practitioners and Researchers, 1305–1314.

Frawley, W. (1984). Prolegomenon to a theory of translation. Translation: Literary, linguistic and philosophical perspectives, 159, 175.

Gellerstam, M. (1986). Translationese in Swedish novels translated from English. Translation studies in Scandinavia, 88–95.

Grabowski, Ł. (2011). Crossing the Frontiers of Linguistic Typology: Lexical Differences and Translation Patterns in English and Russian Lolita by Vladimir Nabokov. In New Perspectives in Language, Discourse and Translation Studies, pp. 227–240. Springer.

Grabowski, Ł. (2013). Quantifying English and Polish Lolitas: A Corpus-Driven Stylistic Comparison. In Correspondences and Contrasts in Foreign Language Pedagogy and Translation Studies, pp. 181–195. Springer.

Hammond, A., Brooke, J., and Hirst, G. (2013). A tale of two cultures: Bringing literary analysis and computational linguistics together. In Proceedings of the 2nd Workshop on Computational Literature for Literature (CLFL13), Atlanta.

Herbelot, A. (2014). The semantics of poetry: A distributional reading. Literary and Linguistic Computing. Hockey, S. (2004). The history of humanities computing. A companion to digital humanities, 3–19.

Hockey, S., and Walker, D. (1993). Developing Effective Resources for Research on Texts: Collecting Texts, Tagging Texts, Cataloguing Texts, Using Texts, and Putting Texts in Context. Literary and Linguistic Computing,8(4), 235–242.

Hung, J.-J., Bingenheimer, M., and Wiles, S. (2010). Quantitative evidence for a hypothesis regarding the attribution of early Buddhist translations. Literary and Linguistic Computing, 25(1), 119–134.

Ilisei, I., and Inkpen, D. (2011). Translationese Traits in Romanian Newspapers: A Machine Learning Approach. International Journal of Computational Linguistics and Applications.

Joachims, T. (1998). Text categorization with support vector machines: Learning with many relevant features. Machine Learning: ECML-98, 137–142.

Jockers, M. L. (2013). Macroanalysis: Digital methods and literary history. University of Illinois Press.

Kenny, D. (2001). Lexis and creativity in translation: a corpus-based study. St Jerome.

Klaussner, C., Lynch, G., and Vogel, C. (2014). Following the trail of source languages in literary translations. In AI-2014: Thirty-fourth SGAI International Conference on Artificial Intelligence, pp. 1–18. Springer.

Koppel, M., and Ordan, N. (2011). Translationese and its dialects. 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, Portland, OR, USA.

Koppel, M., Argamon, S., and Shimoni, A. (2002). Automatically categorizing written texts by author gender. Literary and Linguistic Computing, 17(4), 401–412.

Kurokawa, D., Goutte, C., and Isabelle, P. (2009). Automatic Detection of Translated Text and its Impact on Machine Translation. In Proceedings of the XII MT Summit,Ottawa, Ontario, Canada. AMTA.

Laviosa-Braithwaite, S. (1997). Investigating simplification in an English comparable corpus of newspaper articles. Klaudy and Kohn, 1997, 531–540.

LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature521(7553), 436-444.

Leonardi, V. (2007). Gender and Ideology in Translation: Do Women and Men Translate Differently?: a Contrastive Analysis from Italian Into English, Vol. 301. Peter Lang.

Li, D., Zhang, C., and Liu, K. (2011). Translation Style and Ideology: a Corpus-assisted Analysis of two English Translations of Hongloumeng. Literary and Linguistic Computing, 26(2), 153.

Lucic, A., and Blake, C. (2011). Comparing the Similarities and Differences between Two Translations. In Digital Humanities 2011, p. 174. ALLC.

Lynch, G. (2014). A Supervised Learning Approach Towards Profiling the Preservation of Authorial Style in Literary Translations. In Proceedings of the 25th International Conference on Computational Linguistics (COLING), Dublin, Ireland. Association for Computational Linguistics.

Lynch, G., and Vogel, C. (2009). Chasing the Ghosts of Ibsen: A Computational Stylistic Analysis Of Drama in Translation. In Digital Humanities 2009: University of Maryland, College Park, MD, USA, p. 192. ALLC/ACH.

Lynch, G., and Vogel, C. (2012). Towards the Automatic Detection of the Source Language of a Literary Translation. In Kay, M., and Boitet, C. (Eds.), Proceedings of the 24th International Conference in Computational Linguistics (Coling), Mumbai, India, pp. 775–784. Indian Institute of Technology Bombay.

Manning, C. D., Surdeanu, M., Bauer, J., Finkel, J., Bethard, S. J., and McClosky, D. (2014). The Stanford CoreNLP Natural Language Processing Toolkit. In Proceedings of 52nd Annual Meeting of the Association for Computational Linguistics: System Demonstrations, pp. 55–60 Baltimore, Maryland. Association for Computational Linguistics.

Mazziotta, N. (2010). Building the syntactic reference corpus of medieval French using notabene rdf annotation tool. In Proceedings of the Fourth Linguistic Annotation Workshop, pp. 142–146. Association for Computational Linguistics.

Mendenhall, T. C. (1887). The characteristic curves of composition. Science, pp. 237–246.

Mendenhall, T. C. (1901). A mechanical solution to a literary problem. Popular Science Monthly, 60, 97–105.

Michel, J., Shen, Y., Aiden, A., Veres, A., Gray, M., Pickett, J., Hoiberg, D., Clancy, D., Norvig, P., Orwant, J., et al. (2011). Quantitative analysis of culture using millions of digitized books. Science, 331(6014), 176.

Mikhailov, M., and Villikka, M. (2001). Is there such a thing as a translators style?. In Proceedings of Corpus Linguistics 2001, Lancaster, UK, pp. 378–385.

Mosteller, F., and Wallace, D. L. (1963). Inference in an authorship problem: A comparative study of discrimination methods applied to the authorship of the disputed Federalist Papers. Journal of the American Statistical Association, 58(302), 275–309.

Neuman, Y., Assaf, D., Cohen, Y., Last, M., Argamon, S., Howard, N., and Frieder, O. (2013). Metaphor identification in large texts corpora. PloS One, 8(4), e62343.

Olohan, M. (2001). Spelling out the optionals in translation: a corpus study. UCREL technical papers, 13, 423-432.

Olohan, M. (2002). Comparable corpora in translation research: Overview of recent analyses using the translational English corpus. In LREC Language Resources in Translation Work and Research Workshop Proceedings, pp. 5–9.

Pantopoulos, I. (2009). The stylistic identity of the metapoet: a corpus-based comparative analysis using translations of modern Greek poetry. Ph.D. thesis, The University of Edinburgh.

Popescu, M. (2011). Studying Translationese at the Character Level. In Proceedings of the 8th International Conference on Recent Advances in Natural Language Processing (RANLP’2011). Hissar, Bulgaria.

Remnick, D. (2005). The translation wars. The New Yorker, 7, 98–109.

Rocio, V., Alves, M. A., Lopes, J. G., Xavier, M. F., and Vicente, G. (2003). Automated creation of a Medieval Portuguese partial treebank. In Treebanks, pp. 211–227. Springer.

Rybicki, J. (2006). Burrowing into Translation: Character Idiolects in Henryk Sienkiewicz’s Trilogy and its Two English Translations. Literary and Linguistic Computing, 21(1), 91–103.

Rybicki, J. (2012). The great mystery of the (almost) invisible translator. Quantitative Methods in Corpus-Based Translation Studies: A Practical Guide to Descriptive Translation Research, 231.

Rybicki, J., and Heydel, M. (2013). The stylistics and stylometry of collaborative translation: Woolfs Night and Day in Polish. Literary and Linguistic Computing, 28(4), 708–717.

Saldanha, G. (2011). Translator style: Methodological considerations. The Translator, 17(1), 25–50.

Scott, M. (1996). WordSmith tools.

Selinker, L. (1972). Interlanguage. IRAL-International Review of Applied Linguistics in Language Teaching, 10(1-4), 209–232.

Shlesinger, M., Koppel, M., Ordan, N., and Malkiel, B. (2010). Markers of translator gender: Do they really matter?. Copenhagen studies in language, pp. 183–198.

Stein, A., and Pre´vost, S. (2013). Syntactic annotation of medieval texts. New Methods in Historical Corpora, 3, 275.

Steiner, E. (2002). Grammatical metaphor in translation–some methods for corpus-based investigations1. Language and Computers, 39(1), 213–228.

Vajn, D. (2009). Two-dimensional theory of style in translations: an investigation into the style of literary translations. Ph.D. thesis, University of Birmingham.

van Dalen-Oskam, K. (2012). Names in novels: an experiment in computational stylistics. Literary and Linguistic computing.

van Halteren, H. (2008). Source Language Markers in EUROPARL Translations. In Proceedings of the 22nd International Conference on Computational Linguistics (Coling 2008), pp. 937–944. Association for Computational Linguistics.

Venuti, L. (1995). The translator’s invisibility: A history of translation. Routledge.

Wang, Q., and Li, D. (2012). Looking for translator’s fingerprints: a corpus-based study on Chinese translations of Ulysses. Literary and Linguistic Computing.

Winters, M. (2007). F. Scott Fitzgerald’s Die Schönen und Verdammten: A corpus-based study of speech-act report verbs as a feature of translators’ style. Meta: Journal des traducteurs, 52(3).

Yule, G. U. (1944). The Statistical Study of Literary Vocabulary. Cambridge University Press.

Zanettin, F. (2013). Corpus Methods for Descriptive Translation Studies. Procedia-Social and Behavioral Sciences, 95, 20–32.

Zipf, G. K. (1935). The psycho-biology of language, Houghton Mifflin. Oxford.


[1] Mendenhall mentions De Morgan’s 1872 work A Budget of Paradoxes in the introduction to his 1887 paper, although is himself unsure of the reference location.

[2] A more complete history of computational text analysis in this period is given by Hockey (2004).

[3] Michel et al. (2011) being the canonical work discussing the Google-Ngrams corpus.

[4] See e.g. Candel-Mora and Vargas-Sierra (2013) and Zanettin (2013), who discuss corpus linguistic studies in translation studies literature.

[5] The terms is used to refer to a dialect or variant of a language consisting solely of translations from other languages, and is to some extentrelated to the phenomenon of interlanguage asdefined by Selinker (1972) with reference to thelanguage acquisition context.

[6] See Burrows (2002a) for a detailed explanation of the methodology with examples.

[7] For instance villains vs. heroes, male vs. female.

[8] The data were obtained from the website translationswithoutborders.org.

[9] Examples included readability scores and contractions such as that’s and it’s.

[10] See Altintas et al. (2007) for an example of work on temporally separated Turkish translations.

[11] Google NGrams FAQ estimates POS-tagger accuracy of 90% for 19th century text in their corpus, see http://books. google.com/ngrams/info.

About the author(s)

Dr Gerard Lynch is the Lead Data Scientist at Popertee Ltd, which specialises in location intelligence and campaign management for the new era of engagement. Prior to joining Popertee, he held data science and natural language processing focused roles in the healthtech and media sectors. He received a Phd in Computational Linguistics from Trinity College Dublin in 2013.

Email: [please login or register to view author's email address]

©inTRAlinea & Gerard Lynch (2017).
"Strange bedfellows Shifting paradigms in the corpus-based analyses of literary translations"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2257

Exploration des corpus littéraires à des fins linguistiques et traductionnels

Exemple du connecteur 'mais' en français et de ses variantes traductionnelles en slovène

By Mojca Schlamberger Brezar (University of Ljubljana, Slovenia)

Abstract & Keywords


The paper deals with the distribution of the French connective mais and his translation equivalences in the parallel corpus FraSLoK, composed of French originals and Slovene translations of literary texts, namely novels, and newspaper articles. The connectives and their translations are analysed in order to discover if the type of texte has an impact on the translation. The number of connectives expressing opposition is almost the same as a little bit higher percentage of mais in literary texts whereas there are more variants of French connectives in the part of the corpus consisting of newspaper articles. The distribution of Slovene translation variations in literary and journalistic corpusis slightly different: the analysis reveals the authors’ and translators’ preferences for a certain Slovene connective that as a stylistic variation.


L'article traite les occurences du connecteur français mais et ses variantes traductionnelles dans le corpus parallèle FraSloK, composé d'originaux et de traductions des textes littéraires, notamment des romans, et de textes journalistiques. Les connecteurs et leur traductions sont analysés avec l'intention de voir si le type de texte pourrait influencer leur traduction : en effet, le nombre des connecteurs d’opposition utilisés est à peu près le même avec un taux un peu plus fort de mais en corpus littéraire et un peu plus de variation des connecteurs français en corpus journalistique. La distribution des variantes traductionnelles est pourtant différente dans les deux corpus : l'analyse révèle les affinités de certains auteurs et traducteurs pour un certain connecteur qui est préféré à une autre variante stylistique.

Keywords: corpus littéraires, français, slovène, connecteurs argumentatifs d'opposition, valeur stylistique, literary corpora, French, slovene dialects, argumentative connectives of opposition, stylistic values

©inTRAlinea & Mojca Schlamberger Brezar (2017).
"Exploration des corpus littéraires à des fins linguistiques et traductionnels Exemple du connecteur 'mais' en français et de ses variantes traductionnelles en slovène"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2259

1. Introduction

Notre article propose un regard sur l'exploration de la littérature à des fins linguistiques à la base d'un corpus littéraire parallèle bilingue entre le français en tant que langue source et le slovène en tant que langue de traduction.

Depuis des années et surtout avant l'avènement de la linguistique des corpus, un grand nombre de dictionnaires (monolingues et bilingues) et de grammaires (voir par exemple la Grammaire méthodique : Riegel, Pellat, Rioul 1994) ont puiséleur matériel dans la littérature, aussi bien en slovène qu'en français. En France et dans les pays francophones, on peut constater une émergence des corpus pour le français, accessibles sur l'internet, depuis les années 1990. Avant cette décennie, la seule exploration systématique de la littérature est Frantexte,corpus littéraire français monolingue, et le grand dictionnaire Trésor de la langue française informatisé qui en résulte (TLF dans la suite, http://www.atilf.fr; Habert et al., 1997), qui est un des plus complexes pour la langue française.

En slovène, la tradition de puiser les données linguistiques dans les exemples littéraires pour produire des grammaires et des dictionnaires remonte au début de la tradition de description linguistique de cette langue et est encore valable au XXe siècle : toutes les grammaires sont faites de cette manière, comme le montrent par exemple Toporišič (1976, 2000) ou Vincenot (1975). Le dictionnaire classique de la langue slovène SSKJ était également basé sur des exemples littéraires (Bajec et al., 1994).Selon un de ses éditeurs, F. Jakopin (Bajec et al., 1994 : XVII), les corpus étaient écrits et triés manuellement sur cartes ou fiches – le SSKJ est basé sur 6 millions de fiches et 300 000 mots-clés issus de textes littéraires (en premier lieu !), scientifiques, techniques, journalistiques et autres. Il n’a été informatisé que vers la fin des années 1990.

Depuis toujours, la littérature a réflété la multiplicité des registres de la langue, en permettant également de rentrer ce qui n'est pas « standard » par le biais des dialogues littéraires qui miment l'oral.

Notre article se propose alors l'élaboration des points suivants : dans un premier temps, nous présenterons un regard sur les corpus monolingues et parallèles de la langue slovène, en consacrant une attention spéciale à l'inclusion de la production littéraire. Ensuite, nous donnerons un aperçu du corpus parallèle traductionnel FraSloK, compilé par Adriana Mezeg (Mezeg 2011), et nous offrirons une analyse et comparaison des connecteurs de l'opposition dans les textes littéraires et journalistiques. Les connecteurs d'oppositions se déclinent, en slovène, sur toute une gamme de valeurs stylistiques, que nous voulons exposer sur deux parties du corpus avec l'hypothèse qu'ils vont différer selon les textes explorés.

2. Corpus slovène, attitude des auteurs envers la littérature et la traduction littéraire

Les deux corpus slovènes librement accessibles appartiennent d'un côté la Famille Fida, et de l'autre à la famille Beseda. La 1e, 2e et 3e génération de la famille Fida sont appelées respectivement Fida, Fida Plus et Giga Fida. Le premier corpus Fidaa été compilé entre 1997 et 2000 comme fruit de la collaboration parmi l'Université de Ljubljana, l'Institut de la recherche technologique Jožef Stefan, la maison d'édition DZS et l'entreprise de technologies linguistiques Amebis (http://www.fida.net). A partir de 2007, il a été élargi : le mega-corpus FidaPlus (http://www.fidaplus.net) contient plus de 600 millions de mots. Depuis 2010, un élargissement ultérieur a donnéGigafida (http://www.gigafida.net). Les trois ont été compilés à la base d'articles de journaux, discours du parlement, et surtout de littérature, aussi bien originale que traduite. Le dernier,Gigafida, contient 1,2 milliards de mots provenant de textes sous forme numérique repérés sur Internet ; les genres sont les mêmes que pour les deux autres corpus (http://www.slovenscina.eu/korpusi/gigafida).

Les deux corpus de la famille Beseda, Beseda et Nova Beseda, ont été élaborés dans le cadre de l'institut de la recherche linguistique de l'Académie slovène des arts et des sciences (http://www.bos.zrc-sazu.si/s_beseda.html). Ils sont composés d'œuvres littéraires et d'articles de journauxet comptent quelque 240 millions de mots (Gorjanc, Krek, 2005 : 186-187) ; des traductions, littéraires et pragmatiques, sont présentes.

Cette dernière donnée dit quelque chose sur le statut des traductions en slovène. Le fait d’avoir inclus dans les corpus les traductions va à l'encontre de l'exemple britannique du British National Corpus (BNC), qui semble exclure toute traduction. Ce corpus servait pourtant de modèle aux constructeurs slovènes avec cette reserve que pour le slovène, le statut de la traduction semble donc être comparable au statut de l'original.

Le corpus traductionnel Spook et la partie pour le français et le slovène FraSloK seront présentés dans la suite.

2.1 Le corpus traductionnel SPOOK

A part les corpus monolingues, l'intérêt pour la compilation de corpus traductionnels est apparu en vue de recherches sur les universels en traduction. Le corpus Spook (Slovenski prevodoslovni korpus – Corpus traductionnel slovène),résultat d’un projet dirigé par Špela Vintar[1] entre 2008 et 2012, a été compilé dans le cadre du Département de traduction et interprétation de l’Université de Ljubljana à partir de mai 2009 jusqu’en avril 2012. Il comprend des corpus anglais-slovène, allemand-slovène et italien-slovène, ainsi que la partie français-slovène FraSloK, élaborée par Adriana Mezeg dans le cadre de sa thèse de doctorat et dont nous reparlerons par la suite. Le corpus Spook comprend 95 œuvres littéraires, dont 23 originaux slovènes, le reste étant composé de traductions et de leurs originaux en anglais, italien, allemand et français. Dans le corpus littéraire, seulement les romans sont pris en compte. Déjà dans d’autres corpus (par exemple, Fida et Fidaplus), les auteurs ont inclus surtout de la prose ; étant donné leur taille limitée et leur forte stylisation, la poésie et l'art dramatique ne présentaient pas autant d'intérêt pour leurs compilateurs et sontpartant totalement exclus du corpus.

Le corpus Spook inclut une partie littéraire considérable : comme le note Vintar (2013: 7), l'idée de composition d'origine était beaucoup plus ambitieuse et voulait garantir la représentativité des genres tout en incluant dans la partie non-littéraire des manuels, des articles des journaux et des textes techniques. Mais dans la phase préparatoire, il a été démontré que les traductions non-littéraires actuelles vers le slovène à partir de l’allemand, de l’anglais, du français et de l’italien appartiennent à des genres très différents, qui ne pouvaient pas être comparés entre eux (voir Schlamberger Brezar 2010, Mezeg 2011). La traduction littéraire a donc été choisie comme le seul genre vraiment commun et accessible pour toutes les langues en question.

2. 2 Corpus parallèle FraSloK

La partie française du corpus Spook FraSloK a été compilée par Adriana Mezeg pour les besoins de sa thèse de doctorat intitulée L’analyse des constructions détachées en français et de leurs traductions vers le slovène basée sur les corpus (titre d’origine Korpusno podprta analiza francoskih polstavkov in njihovih prevedkov v slovenščini, Mezeg 2011). Son corpus comprend une partie littéraire composée de 12 romans français et francophones parus entre 1989 et 2007[2], et leurs traductions en slovène dans les 15 dernières années,ainsi qu’une partie journalistique. Il s'agit globalement de 2,5 millions de mots environ (2 466 985 plus exactement). La partie littéraire dépasse un peu la moitié de ces occurrences (1 302 911); quant à la partie journalistique, il s’agit de 300 articles du journal Le Monde diplomatique et de leurs traductions parues en slovène entre 2006 et 2009, avec 1 164 074 occurrences. La partie journalistique est accessible librement à la page internet http://nl.ijs.si/noske tandi,s qu’on peut accéder à la partie littéraire avec un mot de passe[3].

3. Particularités du corpus littéraire

Notre analyse se déroulera sur le corpus des textes littéraires FraSloK. Le corpus est accessible sur le site de l’Institut de technologie Josef Stefan (IJS) et de son Département de technologie cognitive ; il est installé sur le serveur de langue naturelle avec un concordancier et l’outil d’analyse des corpus NoSketch Engine.

Pour cet article, nous avons choisi de faire une recherche quantitative sur l’utilisation des connecteurs de l’opposition, d’abord dans la partie française du corpus littéraire, puis dans la partie slovène qui permettait aussi la recherche parallèle : nous nous sommes concentrée sur mais et ses variantes traductionnelles en slovène. Les connecteurs d'oppositions, très fréquents en français et apparaissant sous formes multiples, sont un enjeu pour le traducteur. A la différence des langues comme l'anglais où il n'y a qu'un ou deux connecteurs d'opposition (but et whereas), le français et le slovène en présentent une variété. Ces connecteurs diffèrent selon leur valeur stylistique, le registre de langue utilisé et présentent en même temps une différence entre l'oral ou l'écrit. Parmi les connecteurs d'opposition en français, mais présente la plus grande variété de valeurs. A la base d'une analyse quantitative, nous voulons établir d'abord sa fréquence par rapport aux autres connecteurs d'opposition, et voir ensuite ses variantes traductionnelles en slovène, en tenant compte des similitudes ou différences dans les parties littéraire et journalistique du corpus FraSloK.

3.1 Les résultats de la recherche quantitative dans les deux corpus 

Dans le corpus littéraire, la totalité de mais était de 2736 occurrences, distribuées dans différents romans comme suit (tableau 1 ci-dessous) :




p | N

Plateforme (2001)


p | N

Le ventre de l'Atlantique (2003)


p | N

Je m'en vais (1999)


p | N

Eldorado (2006)


p | N

Balzac et la petite taill… (2000)


p | N

Truismes (1997)


p | N

L'Amour du prochain (2004)


p | N

Le testament français (1995)


p | N

Mammifères (2003)


p | N

Impératrice (2003)


p | N

Un secret (2004)


p | N

Fou de Vincent (1989)


Tableau 1 fréquence du connecteur mais (2736 occurrences) dans le corpus littéraire

Nous pouvons constater que l’œuvre littéraire qui utilise le plus grand nombre de mais est Plateforme de Michel Houellebecq, tandis que le plus bas nombre se trouve dans le roman Fou de Vincent d’Hervé Guibert.

Les autres connecteurs d’opposition présents dans la partie littéraire du corpus étaient pourtant, cependant, néanmoins. Leur fréquence est inférieure à mais, ils ne représentent même pas 10% des occurrences de mais - comme le montrent les tableaux 2 et 3 ci-dessous.




p | N

Plateforme (2001)


p | N

Truismes (1997)


p | N

Le ventre de l'Atlantique (2003)


p | N

Mammifères (2003)


p | N

Le testament français (1995)


p | N

L'Amour du prochain (2004)


p | N

Je m'en vais (1999)


p | N

Impératrice (2003)


p | N

Balzac et la petite taill… (2000)


p | N

Eldorado (2006)


p | N

Un secret (2004)


p | N

Fou de Vincent (1989)


Tableau 2 : fréquence du connecteur pourtant (199 occurrences dans le corpus littéraire)

Le connecteur pourtant ne comptait que 199 occurrences dans le corpus littéraire mais était utilisé dans tous les romans – de 43 occurrences dans Plateforme jusqu’à trois dans Un secret et Fou de Vincent.




p | N

Je m'en vais (1999)


p | N

Plateforme (2001)


p | N

Mammifères (2003)


p | N

Le testament français (1995)


p | N

Un secret (2004)


p | N

Le ventre de l'Atlantique (2003)


p | N

Impératrice (2003)


p | N

Balzac et la petite taill… (2000)


p | N

Truismes (1997)


p | N

L'Amour du prochain (2004)


p | N

Eldorado (2006)


Tableau 3 : fréquence du connecteur cependant (76 occurrences dans le corpus littéraire)

Le connecteur cependant avec ses 76 occurrences n’était pas présent dans tous les romans (absent du roman Fou de Vincent). Le plus fréquemment employé dans le roman Je m’en vais, il n’atteint qu’une occurrence dans Truismes, L’Amour du prochain et Eldorado.

Les connecteurs néanmoins, avec ses 4 occurrences, et toutefois, avec 12 occurrences, étaient les moins représentés : néanmoins apparaissant dans 3 romans et toutefois dans 5.




p | N

Impératrice (2003)


p | N

Le ventre de l'Atlantique (2003)


p | N

Balzac et la petite taill… (2000)


Tableau 4 : fréquence du connecteur néanmoins (4 occurrences dans le corpus littéraire)




p | N

L'Amour du prochain (2004)


p | N

Le ventre de l'Atlantique (2003)


p | N

Truismes (1997)


p | N

Mammifères (2003)


p | N

Impératrice (2003)


Tableau 5 : fréquence du connecteur toutefois (12 occurrences dans le corpus littéraire)

Pour récapituler, mais est le plus fréquent des connecteurs d'opposition avec ses 2736 occurrences (Tableau 1), pourtant (Tableau 2) apparaît à 199 reprises et les autres connecteurs sont beaucoup moins fréquents – 76 occurrences pour cependant, 4 pour néanmoins et 12 pour toutefois (Tableaux 3, 4 et 5 respectivement).

La recherche quantitative des connecteurs d’oppositions dans le corpus journalistique donne des resultats comparables : mais avec ses 1929 occurrences est toujours le connecteur le plus employé par les journalistes. Toutefois, les autres connecteurs d’opposition sont mieux représentés : pourtant avec 310 occurrences, cependant avec 158, toutefois avec 143 et néanmoins avec 93. A cet égard, le corpus journalistique offre donc une plus grande diversité. Pourtant, mais demeure visiblement le plus important des connecteurs d'opposition dans les deux corpus. Nous présenterons ses valeurs traductionnelles pour le slovène dans la suite de l'article.

3.2 Valeurs du connecteur mais et de ses equivalents traductionnels en slovène

La deuxième recherche portait sur les variantes traductionnelles de mais. Ici, deux points restent à souligner. Le premier est la polysémie du connecteur mais. Le deuxième, l’existence de plusieurs possibilités de traduction de mais vers le slovène, avec différentes valeurs en fonction du registre employé (écrit, oral, soutenu, littéraire, administratif) et du choix de la part du traducteur.

En effet, une multitude de connecteurs slovènes entrent en jeu comme variantes de traduction de mais. Quelles sont les sources de cette diversité ? Le choix des variantes traductionnelles est-il influencé par les dictionnaires, par le processus d’apprentissage de la langue maternelle ou étrangère, par l'acception de la norme commune ou bien est-il tout simplement subjectif? Existe-t-il des stratégies traductionnelles qui permettent de voir les influences possibles sur le choix d'un connecteur spécifique? Nous allons essayer de donner une réponse à ces questions dans la suite.

D’abord, arrêtons-nous à la polysémie du connecteur mais. En tant que connecteur, mais peut prendre plusieurs valeurs, des plus fortement argumentatives, prenant en compte tout le carré argumentatif (voir p. ex. Ducrot etal. 1980 ou Adam 1990), jusqu'aux plus orales, dépouillées de tout environnement argumentatif mais servant à l'argumentation grâce au contexte. Dans l’étude de Jean-Michel Adam (Adam 1990) est présentée la description du connecteur mais à la suite de l’article de O. Ducrot Mais occupe-toi d’Amélie (Ducrot et al. 1980) où sont exposées cinq valeurs :

Mais réfutatif (Adam 1990 : 194), construit sur une négation et combinable avec non pas, plus, point... Il peut être présenté à l’aide de l’exemple : ce n’était pas lui qui parlait mais sa luxure...

Mais concessif, qui apparaît toujours en connexion avec la négation, même si celle-ci est implicite, p. ex. Cet étudiant est intelligent, mais il échoue à tous les examens. Le mais concessif mène alors à une conclusion directe. Le mais réfutatif se construit avec la négation de la première proposition qui se combine avec non pas (plus, point), selon le schéma non pas P, mais Q (Adam 1990: 194), comme par exemple dans Il n'était pas malade, mais paresseux. Ici, l’enchaînement à l’opposition à la première partie de l’énoncé se fait directement et il n’est pas nécessaire d’y accéder à travers des conclusions explicites ou implicites.

Mais argumentatif, le plus complexe de tous. Il comprend deux arguments explicites et deux conclusions qui sont le plus souvent implicites mais toujours en opposition (Adam 1990: 206), comme dans « Cet étudiant est intelligent, mais paresseux ». Le mais argumentatif est basé sur les présuppositions des locuteurs concernant l’état des choses et leur attitude envers le contexte et les normes civilisationnelles, ce que Ducrot et Anscombre (Anscombre, Ducrot 1983) ont dénommé le topos. Dans ce cas-là, la qualité P (être intelligent) n’implique pas non-Q (ne pas être paresseux) : cette implicature n’est pas basée sur le monde extérieur mais sur les représentations et les présuppositions du locuteur à propos du monde qui l’entoure. Le mais argumentatif fait partie de l’espace discursif où P présente l’argument pour la conclusion C et Q l’argument pour la conclusion non-C, et où mais est l’indicateur que la conclusion est argumentativement orientée vers non-C.

Les différences entre ces trois mais sont minimes aussi bien dans leur contenu conceptuel que dans les variantes en traduction, qui se jouent sur les mêmes équivalents slovènes, le plus fréquent et le plus proche étant le connecteur ampak.

Nous traîterons à part le mais de renforcement-renchérissement et le mais phatique. Le mais de renforcement-rencherissement est le connecteur qui apparaît dans des exemples tels que non seulement … mais aussi (Adam 1990 : 192). Ce type de proposition est généralement construit avec non seulement dans la proposition P et combiné avec même, aussi, également, en plus, de plus dans la proposition Q. C’est à dire qu’il apporte un argument supplémentaire pour une conclusion qui peut être exprimée explicitement ou implicitement. Ce mais donne généralement en slovène « ne samo, ampak / marveč / temveč tudi ».

L’autre valeur de mais, qui peut être traîtée à part, est celle du mais phatique, qui apporte la démarcation des segments textuels (Adam 1990 : 197, Schlamberger Brezar 2009 : 205-207) et signale une sorte d’opposition dérivée d’une forme orale comme dans Mais faites donc attention! (voir aussi Schlamberger Brezar 2012). Nous pouvons expliquer son fonctionnement en le considérant comme un mais de l’oral qui garde sa valeur argumentative mais apparaît dans des propositions elliptiques, agit en fonction du contexte implicite et se combine avec les marqueurs discursifs, par exemple mais oui, mais bien sûr, mais quand même. Ses valeurs dépendent du contexte exclamatif, impératif ou interrogatif mais il peut aussi acquérir une fonction métalinguistique (Schlamberger Brezar 2009 : 206). Ce type de connecteur d’opposition apparaît le plus souvent dans les dialogues où il serait traduit vers le slovène par ampak ou pa, les deux étant les marqueurs de l’oral par excellence.

Comme nous l'avons déjà annoncé dans la partie précédente, les equivalents ou les homologues de mais en traduction vers le slovène sont nombreux, le plus proche étant ampak. C’est pourtant une question de normativité, de choix personnel et de contraintes textuelles, entre autres. Nous citons dans la suite les possibilités de choix des connecteurs données par les dictionnaires et les grammaires.

Dans le dictionnaire français-slovène Francosko-slovenski slovar (Grad 1984) les traductions slovènes suivantes sont données pour mais : conjonction toda, ampak; temveč ; non seulement mais encore : ne le, temveč tudi. Claude Vincenot dans son Essai de grammaire slovène (Vincenot 1975: 297) à l’intention des locuteurs français parle des « jonctifs adversatifs simples » où il énumère a/toda/ampak/ marveč/ temveč = mais, a/ali (littéraire) =mais, vendar = pourtant, cependant, « jonctifs adversatifs doubles qui sont concessifs » sicer, vendar – certes, pourtant ainsi que les conjonctions marquant le « renchérissement positif » : ne samo/ le … ampak, marveč, temveč tudi (non seulement, mais encore). D’après Vincenot, les équivalents de mais les plus appropriés seraient pa, toda, ampak, marveč, temveč, un peu moins vendar et samo.

Cette détermination des connecteurs dans le dictionnaire offre au traducteur des informations sur la polysémie et la nécessité de prendre en compte le contexte minimal, donnant tout de même une multitude de possibilités traductionnelles. Le dictionnaire de Grad et la grammaire de Vincenot fournissent le minimum d’indications aux traducteurs pour un choix adéquat de l’équivalent mais vont aussi vers la synonymie des connecteurs.

Selon le grand dictionnaire slovène monolingue Slovar slovenskega knjižnega jezika (dorénavant SSKJ, Bajec et al. 1994), les connecteurs d’opposition sont interchangeables, chacun possédant cependant une particularité stylistique. La conjonction a est qualifiée de plutôt littéraire, possible dans la coordination copulative. Pour ampak, les valeurs suivantes sont citées : 1) dans la coordination adversative, 2) valeur expressive ne samo…ampak tudi (non seulement... mais aussi qui correspond au mais de renchérissement), 3) expressif pour l’opposition ; 4) expressif, oral, au début de la phrase. Ici, ampak correspond au mais phatique. Les valeurs du connecteur pa – conjonction dans la coordination adversative, pour exprimer une faible opposition – sont typiques de l’oral. Les conjonctions et connecteurs toda et vendar partagent la possibilité d’être employées en coordination adversative, mais expriment aussi l’opposition. Le connecteur toda est plutôt littéraire et soutenu, vendar ajoute la valeur qu’une certaine conclusion va de soi. Les adverbes adversatifs marveč et temveč donnent plus d’impact à la force d’expression des arguments et peuvent aussi accompagner ampak en tant que traductions du mais de recherissement.

Ces désignations dans le dictionnaire pourraient montrer le fil conducteur du comportement des traducteurs. Dans le corpus littéraire, les passages diffèrent en approche narrative, descriptive ou dialogale. Les dialogues sont souvent la source de l’oralité, alors nous supposons que le traducteur proposera des traductions différentes selon le prototype de la séquence (Schlamberger Brezar, 2012). En même temps, un certain poids de la subjectivité des traducteurs dans le choix des connecteurs slovènes qui à leur tour présentent l'équivalence de mais en slovène est à prévoir.

3.3 Les variantes traductionnelles de mais et leur fréquence dans le corpus littéraire

En corpus parallèle littéraire, nous avons trouvé comme la variante traductionnelle de mais les connecteurs vendar, pa, ampak, a, toda, marveč et temveč. Nous fournirons les résultats de notre analyse en tenant compte de leur fréquence décroissante, tout en évaluant les choix des traducteurs. Nous avons évalué les choix des traducteurs pour différentes stratégies, basées sur le style de l'auteur ou sur leur propre poétique ou politique linguistique, ou parfois suggérés aussi par les relecteurs des documents.

Les résultats quantitatifs de la traduction des connecteurs vers le slovène dans les traductions des textes littéraires sont les suivants. La variante traductionnelle du connecteur mais la plus fréquemment utilisée était le connecteur vendar avec ses 766 occurrences :




p | N

Eldorado (2007)


p | N

Francoski testament (2005)


p | N

Platforma (2002)


p | N

Cesarica (2007)


p | N

Balzac in kitajska šivilj… (2002)


p | N

Sesalci (2007)


p | N

Ljubezen do bližnjega (2007)


p | N

Trebuh Atlantika (2007)


p | N

Grem (2008)


p | N

Skrivnost (2008)


p | N

Svinjarije (1997)


p | N

Nor na Vincenta (2007)


Tableau 6 : fréquence du connecteur vendar (766 occurrences dans le corpus littéraire)

Vendar est le moins fréquent dans la traduction du roman Fou de Vincent et il semble être le connecteur préféré de la traductrice du roman Eldorado. Cette traduction semble automatique : en effet, la traductrice n’utilise que très peu les autres connecteurs slovènes en combinaison avec mais, pa (18 occurrences), ampak (12 occurrences), a (9), temveč (2), marveč (3) et aucun toda. Suit le connecteur pa :




p | N

Balzac in kitajska šivilj… (2002)


p | N

Platforma (2002)


p | N

Svinjarije (1997)


p | N

Trebuh Atlantika (2007)


p | N

Grem (2008)


p | N

Ljubezen do bližnjega (2007)


p | N

Francoski testament (2005)


p | N

Sesalci (2007)


p | N

Cesarica (2007)


p | N

Eldorado (2007)


p | N

Nor na Vincenta (2007)


p | N

Skrivnost (2008)


Tableau 7 : fréquence du connecteur pa (602 occurrences dans le corpus littéraire)

Le connecteur pa oral est moins contraignant en adversativité que ampak (plus faible) mais aussi appartenant au registre oral et à la traduction dans les dialogues. Ce connecteur est le plus fréquemment utilisé dans la traduction du roman Balzac et la petite tailleuse chinoise (Balzac in mala kitajska šiviljica) avec 94 occurrences, suivi de très près de Plateforme – Platforma (90 occurrences). Il est intéressant de voir que le nombre des occurrences reste assez élevé dans les quatre romans qui suivent (de 75 à 59). Ce connecteur est le moins fréquent dans la traduction du roman Un secret. Le connecteur pa, fortement oral en slovène, entre en concurrence avec ampak, discuté dans le tableau 8. Il paraît que les deux ont une fréquence faible dans les œuvres présentant le mois de dialogues (SkrivnostLe Secret, 7 pa contre 9 ampak ; SvinjarijeLes Truismes, 75 pa contre 164 ampak etc.):




p | N

Svinjarije (1997)


p | N

Sesalci (2007)


p | N

Platforma (2002)


p | N

Grem (2008)


p | N

Balzac in kitajska šivilj… (2002)


p | N

Ljubezen do bližnjega (2007)


p | N

Trebuh Atlantika (2007)


p | N

Francoski testament (2005)


p | N

Eldorado (2007)


p | N

Skrivnost (2008)


p | N

Nor na Vincenta (2007)


p | N

Cesarica (2007)


Tableau 8 : fréquence du connecteur ampak (583 occurrences dans le corpus littéraire)

Le plus grand nombre de mais traduits par ampak apparaît dans la traduction du roman Truismes. Le nombre total des occurrences de mais dans ce roman était de 229, dont 164 sont traduites par ampak. Des autres traductions possibles de mais apparaîssent toda (8 occurrences), pa (75 – ce qui prédit l’oralité de l’oeuvre), et vendar (4), avec une absence totale de a, marveč ou temveč. Le choix du connecteur ampak dans ce roman est bien fondé, puisqu’il s’agit d’une narration un peu naïve en 1e personne. Il en va de même pour le choix des connecteurs dans la traduction des Mammifères – Sesalci de Pierre Mérot (114 occurrences de ampak) où le cadre de la narration est presque pareil à Truismes – Svinjarije, sauf qu’il se déroule avec un peu de distance, un narrateur omniprésent racontant l’histoire de l’oncle en 3e personne, mais la narration est très oralisée. Il parait que le traducteur en a tenu compte.

Le connecteur toda du Tableau 9, avec la connotation «littéraire, soutenu » est assez fréquent dans trois romans comme variante de mais : Platforme (81), Ventre de l’Atlantique (78) et L’impératrice (73). Le moins fréquent est dans le roman Je m’en vais où prédomine le connecteur a qui a la même valeur (voir le Tableau10).

p | N

Platforma (2002)


p | N

Trebuh Atlantika (2007)


p | N

Cesarica (2007)


p | N

Skrivnost (2008)


p | N

Ljubezen do bližnjega (2007)


p | N

Nor na Vincenta (2007)


p | N

Svinjarije (1997)


p | N

Sesalci (2007)


p | N

Balzac in kitajska šivilj… (2002)


p | N

Grem (2008)


p | N

Francoski testament (2005)


Tableau 9 : fréquence du connecteur toda (324 occurrences dans le corpus littéraire)

Le connecteur a, synonyme de toda selon la stylistique (avec sa valeur soutenue et littéraire) compte, en tant que variante traductionnelle de mais, 437 occurrences. Ce connecteur n’apparaît pas dans toutes les traductions (absent dans Truismes et Testament français). On en trouve le plus grand nombre dans la traduction du roman Je m’en vais, à savoir 191 occurrences. Les connecteurs toda et a présentent la même valeur stylistique et une fréquence comparable, avec cette reserve que le choix de toda élimine le choix de : la traduction Skrivnost – Le Secret présente 49 occurrences de toda mais une seule de a ; le connecteur a apparaît 191 fois dans Grem – Je m'en vais, où toda est complètement absent. Le connecteur a s’est alors montré être la préférence subjective de la traductrice de Grem – Je m'en vais, et la distribution de toda ou a semble être liée à un choix subjectif des traducteurs.




p | N

Grem (2008)


p | N

Trebuh Atlantika (2007)


p | N

Ljubezen do bližnjega (2007)


p | N

Platforma (2002)


p | N

Cesarica (2007)


p | N

Nor na Vincenta (2007)


p | N

Balzac in kitajska šivilj… (2002)


p | N

Eldorado (2007)


p | N

Sesalci (2007)


p | N

Skrivnost (2008)


Tableau 10 : fréquence du connecteur a (437 occurrences dans le corpus littéraire)

Dans les prochains tableaux figurent les traductions du mais de renchérissement sous forme des connecteurs slovènes temveč (Tableau 11), marveč (Tableau 12) et leur variante moins soutenue ampak tudi (Tableau 13).

Le connecteur temveč est le plus fréquent dans la traduction du roman Le Ventre de l’Atlantique mais complètement absent des traductions des romans Je m’en vais, Truismes, Mammifères, Balzac et la petite tailleuse chinoise et Testament français où il semble être remplacé par la variante moins littéraire, plus moderne ampak tudi. Par contre, marveč n’obtient que 5 occurrences dans trois romans, Eldorado, Platforme et L’Amour du prochain. Dans la traduction de ce dernier, il existe une occurrence pour chaque connecteur susceptible de traduire un mais de renchérissement ; marveč, temveč et ampak tudi.




p | N

Trebuh Atlantika (2007)


p | N

Platforma (2002)


p | N

Cesarica (2007)


p | N

Nor na Vincenta (2007)


p | N

Eldorado (2007)


p | N

Skrivnost (2008)


p | N

Ljubezen do bližnjega (2007)


Tableau 11 : fréquence du connecteur temveč (38 occurrences dans le corpus littéraire)




p | N

Eldorado (2007)


p | N

Platforma (2002)


p | N

Ljubezen do bližnjega (2007)


Tableau 12 : fréquence du connecteur marveč (5 occurrences dans le corpus littéraire)




p | N

Balzac in kitajska šivilj… (2002)


p | N

Svinjarije (1997)


p | N

Sesalci (2007)


p | N

Platforma (2002)


p | N

Ljubezen do bližnjega (2007)


Tableau 13 : fréquence du connecteur ampak tudi (11 occurrences dans le corpus littéraire)

Avant de conclure, nous pouvons comparer les résultats obtenus de l’analyse quantitative du corpus littéraire avec ceux du corpus journalistique où les chiffres sont les suivants :[4]le connecteur le plus fréquent pour mais français dans la partie journalistique est vendar, qui compte 541 occurrences. Suit pa avec 522, toda avec 304, a avec 256, ampak avec 204 et temveč avec 82 occurrences (ce dernier en concurrence avec ampak tudi, apparaissant 28 fois). Le connecteur marveč, pourtant repérédans le corpus littéraire, n’apparaît pas dans le corpus journalistique. La plus grande différence entre les deux corpus apparaît dans le cas de la traduction de mais avec le connecteur ampak, qui compte 583 occurrences dans le corpus littéraire et seulement 204 dans la partie journalistique du FraSloK. Comme ce connecteur est censé appartenir au registre oral, on peut faire l’hypothèse d’une (auto)censure ou simplement d’une intervention des rédacteurs du texte[5]. Pour cette conclusion, nous nous basons sur le fait que ampak était le plus fréquemment employé dans les œuvres littéraires qui reproduisaient l'oral.

Il en va de même pour temveč, qui est de 50% plus fréquent en corpus journalistique (82 occurrences) qu'en corpus littéraire (38 occurrences). Mais finalement, les données de fréquence dans les deux corpus correspondent en gros aux résultats obtenus dans les corpus monolingues.[6]

4. Conclusion

La comparaison entre la partie littéraire et journalistique du corpus bilingue parallèle français-slovène offre des résultats intéressants : le nombre des connecteurs d’opposition utilisés est à peu près le même avec un taux un peu plus fort de mais en corpus littéraire et un peu plus de variation des connecteurs en corpus journalistique. A partir de ce résultat, on pourrait supposer plus d’argumentativité dans la partie journalistique et plus de mimésis d’oralité dans la partie littéraire.

Grâce à l’analyse quantitative des connecteurs, nous avons accès à une certaine poétique de l’auteur – parmi les romans inclus dans le corpus le plus grand nombre de connecteurs d’oppositions est inclus dans Plateforme de Michel Houellebecq. Le choix des équivalents de mais en slovène se fait selon le registre de la langue (soutenu, familier), le type de texte (narratif ou dialogique) et les prédilections subjectives des traducteurs ; certains préfèrent a, d’autres toda ou vendar. Ainsi, la poétique d’un traducteur particulier peut ressortir : on peut constater que certains auteurs ont leurs préférences concernant le choix de la variante traductionnelle de mais : par exemple, ampak est le connecteur préféré dans les Truismes – ce choix peut être expliqué par la nature de ce roman qui est raconté en 1e personne et alors fortement sous l’impact de l’oral. Il est suivi de très près des Mammifères reprenant la même stylistique mais à la 3e personne. Par contre, vendar apparaît presque comme la seule variante traductionnelle de mais dans l’Eldorado où le traducteur donne l’impression de choisir ce connecteur automatiquement même si ce choix n’est pas toujours le meilleur. On peut aussi constater que le connecteur a présente la prédilection de la traductrice du roman Je m’en vais. Il est intéressant que dans le corpus FraSloK, où les romans inclus datent des 30 dernières années et les traductions de 15, quelques traducteurs respectent ces tendances stylistiques de modernité en se servant des connecteurs qui, d’après le dictionnaire SSKJ sont censés être oraux (ampak, pa), tandis que d’autres cherchent à établir la littérarité avec l’emploi des connecteurs plus soutenus (toda, a).

Nous pouvons conclure que les corpus littéraires, avec leur multitude de registres linguistiques inclus dans différents procédés textuels, représentent une langue plus riche, plus variée que les corpus faits de textes spécialisés, par exemple le corpus journalistique – et que ce fait, même s’il ne présente pas la réalité linguistique, pourrait au moins représenter un idéal vers lequel devraient s'élever les locuteurs natifs et les apprenants étrangers d'une langue.


Adam, Jean-Michel (1990). Eléments de linguistique textuelle. Bruxelles : Mardaga.

Adam, Jean-Michel (1994). Les textes : types et prototypes. Paris : Nathan.

Anscombre, Jean-Claude, Ducrot, Oswald (1983). Argumentation dans la langue. Bruxelles : Mardaga.

Bajec, Anton et al. : Slovar slovenskega knjižnega jezika. Ljubljana : DZS.

Baker, Mona (1995). « Corpora in translation studies. An overview and some suggestions for future research. » In Target 7 223-243.

Ducrot, Oswald et al. (1980). Les mots du discours. Paris : Minuit.

Gorjanc, Vojko (2005). Uvod v korpusno jezikoslovje. Domžale : Izolit.

Gorjanc, Vojko, Krek Simon (éd.) (2005). Študije o korpusnem jezikoslovju. Ljubljana : Krt.

Grad, Anton (1984). Francosko-slovenski slovar. Ljubljana : DZS.

Habert, Benoît, Nazarenko. Adeline, Salem, André (1997). Les linguistiques de corpus. Paris : Armand Colin.

Kennedy, Graeme (1998). An introduction to Corpus Linguistics. London et New York: Longman.

Mezeg, Adriana (2010). « Compiling and using a French-Slovenian parallel corpus. » In Xiao, Richard (éd.). Proceedings of The International Symposium on Using Corpora in Contrastive and Translation Studies : 2010 conference (UCCTS2010). Ormskirk : Edge Hill University<http://www.lancs.ac.uk/fass/projects/corpus/UCCTS2010Proceedings/papers/Mezeg.pdf>, accès mars-mai 2014.

Mezeg, Adriana (2011). Korpusno podprta analiza francoskih polstavkov in njihovih prevedkov v slovenščini : thèse du doctorat. Ljubljana: Univerza v Ljubljani.

Olohan, Maeve (2004). The use of corpora in translation studies. London-New York Routledge.

Schlamberger Brezar, Mojca (2000). « Les connecteurs en combinaison avec lesmarqueurs modaux : l'exemple du français et du slovène », In Linguistica 40 273-282.

Schlamberger Brezar, Mojca (2002). « Le rôle syntaxique et pragmatiquedes connecteurs dans le discours argumentatif français », In Linguistica 42 89-110.

Schlamberger Brezar, Mojca (2005). « Mais - inverseur d'orientation argumentative. Mais justement? ». In Vilkou-Poustovaia, Irina (éd.), L'exception entre les théories linguistiques et l'expérience, Faits de langues 25, Paris, Gap : Ophrys 146-150.

Schlamberger Brezar, Mojca (2009). Povezovalci v francoščini: od teoretičnih izhodišč do vloge v diskurzu. Ljubljana : Znanstvena založba Filozofske fakultete.

Schlamberger Brezar, Mojca (2010). « La préparation des corpus parallèles et comparables et leur utilité pour le traducteur », In Cahiers du GEPE Traduction et domaines de spécialités (Actes des Journées d’études de Strasbourg « Outils de traduction, outils du traducteur ? », Université Marc-Bloch- Strasbourg 2, 18-19 septembre 2008), <http://www.cahiersdugepe.fr/index.php?id=1654#tocto1n1> (accès septembre 2014)

Schlamberger Brezar, Mojca (2011). « Les marqueurs de l'oral en français et en slovène et la justification de leur utilisation en interprétation.» Pavelin Lešić, Bogdanka (éd.). Francontraste 2. La francophonie comme vecteur du transculturel. Mons : CIPA 255-262.

Schlamberger Brezar, Mojca (2012). « Les marqueurs discursifs « mais » et «alors » en tant qu'indicateurs du degré de l'oralité dans les discours officiels, les débats télévisés et les dialogues littéraires ». Linguistica 52, 225-237.

Schlamberger Brezar, Mojca (2013). « Primerjava slovenskih prevodnih variant povezovalca mais v francosko-slovenskem vzporednem korpusu. » Vintar, Špela (éd.): Slovenski prevodi skozi korpusno prizmo, Ljubljana : Znanstvena založba Filozofske fakultete.

Toporišič, Jože (1976, 2000). Slovenska slovnica. Maribor : Obzorja.

Vincenot, Claude (1975). Essai de grammaire slovène. Ljubljana : Mladinska knjiga.

Vintar, Špela (2013). Slovenski prevodi skozi korpusno prizmo. Ljubljana: Znanstvena založba Filozofske fakultete.



Beseda, Nova Beseda: www.bos.zrc-sazu.si/s_beseda.html (accès juillet-septembre 2014)

Fida plus: www.fidaplus.net (accès juillet-septembre 2014)

FraSloK: http://nl.ijs.si/noske (accès juillet-septembre 2014)

GigaFida: http://www.gigafida.net (accès juillet-septembre 2014, juillet 2016) [

SPOOK: http://nl.ijs.si/noske (accès juillet-septembre 2014)

Trésor de la Langue Française informatisé TLFi: http://atilf.atilf.fr (accès juillet-septembre 2014) 



[1] Špela Vintar, le projet J6-2009, Slovensko prevodoslovje – viri in raziskave (la traductologie slovène – ressources et recherche).

[2] Les romans originaux et leurs traductions (entre parenthèses) sont les suivants : Fou de Vincent de Hervé Guibert (Nor na Vincenta, traduction de Brane Mozetič), Un Secret de Philippe Grimbert (Skrivnost, traduction de Jožica Grum), Impératrice de Sa Shan (Cesarica, traduction de Ana Barič), Eldorado de Laurent Gaudé (Eldorado, traduction de Tjaša Mohar), Balzac et la petite tailleuse chinoise de Dai Sijie (Balzac in kitajska šiviljica, traduction de Alenka Moder Saje), Testament français d'Andreï Makine (Francoski testament, traduction de Nadja Dobnik), Amour du prochain de Pascal Bruckner (Ljubezen do bližnjega, traduction de Jaroslav Skrušny), Je m'en vais de Jean Echnoz (Grem, traduction de Suzana Koncut), Mammifères de Pierre Mérot (Sesalci, traduction de Jan Jona Javoršek), Le Ventre de l'Atlantique de Fatou Diome (Trebuh Atlantika, traduction de Ana Barič), Truismesde Marie Darrieusecq (Svinjarije, traduction de Marko Crnkovič), Plateforme de Michel Houellebecq (Platforma, traduction de Mojca Medvedšek.

[3] Le problème des droits d’auteuts n’ayant pas été résolu pour toutes les langues, le corpus n’est destiné qu’aux chercheurs qui ont participé au projet de Špela Vintar.

[4] Ces chiffres sont présentés en dehors du tableau par les limites spatiales.

[5] Ce genre d'interventions est assez fréquent en slovène aussi bien dans les textes littéraires qu'autres.

[6]Gigafida présente 662014 occurrences de ampak, 446531 occurrences de toda, 218994 occurrences de temveč et 22458 occurrences de marveč.

About the author(s)

Mojca Schlamberger Brezar is a full professor of French linguistics and translation at the Department of Translation Studies at the University of Ljubljana. She has a PhD in argumentative connectives in French spoken discours. Her main interests are contrastive grammar and translation between French and Slovene.

Email: [please login or register to view author's email address]

©inTRAlinea & Mojca Schlamberger Brezar (2017).
"Exploration des corpus littéraires à des fins linguistiques et traductionnels Exemple du connecteur 'mais' en français et de ses variantes traductionnelles en slovène"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2259

Participe passé détaché dans des romans contemporains français et leurs traductions slovènes

By Adriana Mezeg (University of Ljubljana, Slovenia)

Abstract & Keywords


This article deals with the use of past participles in French non-finite clauses and their translation into Slovenian. We shall focus on sentence-initial participial clauses as these are the most difficult to understand and to decode for a non-native French speaker/learner. Having a single form for three different functions, we distinguish three categories of a French past participle of verbal origin because we expect that due to their different functioning, this might influence their translation into Slovenian. The study is based on the examples extracted semi-automatically from the literary part of the FraSloK parallel corpus which comprises twelve contemporary French novels and their Slovenian translations. The present article proposes to provide answers to some questions important for the fields such as corpus linguistics, French-Slovenian contrastive syntax and translation studies, for instance about the use and utility of parallel corpora in studying complex structures such as participial clauses, about the syntactic and semantic differences/similarities between French and Slovenian literary texts, and about the translation strategies revealed, useful in the pedagogic as well as professional context. 


Cet article aborde l’emploi des participes passés dans des constructions détachées françaises et leur traduction vers le slovène. Nous ne traiterons que les structures détachées en position initiale, car celles-ci s’avèrent le plus difficiles à comprendre et à décoder pour un locuteur / apprenant francophone non natif. Possédant une seule forme pour trois fonctions différentes, nous distinguons trois catégories de participe passé français d’origine verbale, supposant que, en raison de leur fonctionnement différent, cela puisse influer sur leur traduction vers le slovène. L’étude porte sur les exemples extraits semi-automatiquement de la partie littéraire du corpus parallèle FraSloK, contenant douze romans contemporains français et leurs traductions slovènes. L’article se propose de répondre à quelques questions importantes pour les domaines comme la linguistique de corpus, la syntaxe contrastive franco-slovène et la traductologie, par exemple sur l’emploi et l’utilité des corpus parallèles pour la recherche des structures complexes telles que les constructions détachées, sur les différences / similitudes syntaxiques et sémantiques entre les textes littéraires français et slovènes, et sur les stratégies traductionnelles décelées, utiles dans le contexte pédagogique de même que professionnel.  

Keywords: participe passé, construction détachée, traduction, explicitation, corpus parallèle, past participle, participial clause, parallel corpus

©inTRAlinea & Adriana Mezeg (2017).
"Participe passé détaché dans des romans contemporains français et leurs traductions slovènes"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2256

1. Introduction

Dans la langue française, lorsqu’on parle du participe passé, l’une des premières choses qui nous viennent à l’esprit est qu’il sert à former les temps composés, par exemple :

(1) Jacques est arrivé à la maison.

(2) Stéphane a regardé le match.

Or, le participe passé est une forme complexe qui peut avoir différentes valeurs et fonctions syntaxiques et figurer dans différentes structures de la phrase. Dans cette étude, nous abordons les participes passés français employés sans auxiliaire dans les structures que nous appelons, d’après Combettes (1998), détachées. Ses caractéristiques principales sont : a) la séparation du reste de la phrase par une ponctuation non finale à l’écrit ; b) la prédication seconde, ce qui veut dire que le prédicat (sous forme d’un gérondif, d’un participe présent ou d’un participe passé)[1] instaure avec le sujet (plus rarement avec l’objet ou un autre actant) de la proposition principale une relation identique à celle dans la prédication principale ; c) la relation de coréférence entre le prédicat de la phrase détachée et le sujet de la proposition principale (Combettes 1998 : 10-13). Nous ne nous concentrons que sur les structures ayant comme noyau un participe passé et figurant au tout début d’une phrase, car celles-ci semblent être le plus difficiles à décoder pour un locuteur francophone non natif. Prenons un exemple :

(3) Assise au bord du lac, Marie réfléchit sur son futur.

Ce qui pose problème pour un locuteur slovène au niveau de la compréhension et de la traduction du français vers le slovène, c’est le fait que le sujet n’est pas explicitement indiqué (par exemple sous forme d’un pronom ou d’un groupe nominal) dans la structure de surface et que la relation sémantique entre la structure détachée et la proposition principale (par exemple temporelle ou explicative, comme dans (3)) reste implicite en raison des connecteurs non-exprimés. En outre, cette relation est parfois difficile à établir (comme dans (3)), ce qui nous oblige à prendre en considération le contexte le plus large et à utiliser tout notre savoir (extra)linguistique pour arriver à une interprétation correcte.

Bien que la langue slovène connaisse les structures détachées, elles semblent être d’un emploi rare[2] et ont l’air archaïque dans les textes écrits contemporains, ainsi supposons-nous que lors de la traduction du français, nous avons tendance à les exprimer en slovène par d’autres structures syntaxiquement et sémantiquement plus explicites, plutôt qu’à les garder. En effet, selon Breznik (1967 : 82-94), il s’agirait d’une structure romane entrée dans la langue slovène par la traduction d’ouvrages littéraires français du début du XXème siècle (par exemple écrits par Flaubert, Zola, Anatole France). D’après lui (1967 : 94), la structure détachée est peu naturelle dans la syntaxe slovène où il est primordial que toute phrase contienne un verbe.[3] Ainsi sera-t-il intéressant de vérifier quel est le taux de structures détachées dans les textes sources français et dans les traductions slovènes de notre corpus.

Le présent article comporte, dans un premier temps, une partie théorique où nous discutons le participe passé français et proposons trois types selon leurs fonctions, qui peuvent former une structure détachée : le participe passé qui, à l’actif, désigne une action, le participe passif et le participe d’état. Cette distinction nous aidera lors de la classification des exemples sources et cibles et nous permettra à discerner les différences et similitudes dans les traductions slovènes. Ensuite, nous présentons le corpus de travail, constitué de douze romans contemporains français et de leurs traductions slovènes, ainsi que le processus et les résultats de l’extraction semi-automatique des structures détachées comportant un participe passé du corpus en question.

Dans un second temps, nous analysons les traductions slovènes du point de vue syntaxique et sémantique pour voir comment les traducteurs ont rendu en slovène les structures détachées françaises. Nous concluons l’article en essayant de répondre aux questions que soulève cette recherche, notamment sur l’emploi et l’utilité des corpus parallèles pour la recherche des structures complexes telles que les constructions détachées, sur les différences / similitudes syntaxiques et sémantiques entre les textes littéraires français et slovènes, et sur les stratégies traductionnelles révélées, utiles dans le contexte pédagogique de même que professionnel.

2. Le participe passé en français

Comme nous l’avons mentionné dans la partie introductoire, une construction détachée peut se former autour d’un gérondif, un participe présent, un participe passé, un nom ou groupe nominal, ou bien un adjectif. Pour les fins de cette étude, nous avons décidé de nous pencher sur les structures comportant un participe passé, car cette forme grammaticale se révèle intéressante pour ses différentes valeurs et fonctions, comme nous allons voir dans la suite.

En français, le participe passé possède deux formes :

– une forme simple : elle est constituée du radical verbal auquel on ajoute différentes désinences : pour les verbes du 1er groupe, -i pour ceux du 2ème groupe et , -i, -u, -t, -s et -ert pour les verbes du 3ème groupe. Dans une structure détachée, la forme simple du participe passé peut être précédée par un adverbe (comme trop, très), une conjonction (comme mais) ou un pronom (comme lui) et s’accorde avec le sujet de la proposition principale en genre et en nombre ;

– une forme composée : elle est constituée du participe présent du verbe avoir (ayant) ou être (étant) et de la forme simple du participe passé (par exemple ayant trouvé, étant sorti(e)(s)). Dans le cas du verbe être, le participe passé s’accorde avec le sujet en genre et en nombre ; on l’utilise avec les verbes perfectifs qui expriment un mouvement (par exemple ayant fini, étant sorti) ou avec les verbes pronominaux (par exemple m’étant souvenu) (Riegel, Pellat et Rioul 1999 :252). La forme composée du participe passé peut être précédée par un adverbe (comme ne, puis), une conjonction (par exemple et, mais) ou un pronom (comme se, me, le). Elle exprime une action accomplie, l’antériorité par rapport au verbe principal.

Dans notre recherche, nous prendrons en considération tous les participes passés de notre corpus ayant une origine verbale, sinon on pourrait classifier certains parmi les constructions détachées adjectivales, car dans une structure détachée, ils peuvent exprimer l’état après une action accomplie ou une caractéristique, et non une action, par exemple :

(4) Pierre a surpris Tania. (action) / Surprise, Tania ne savait que répondre. (état)

Cela nous paraît important pour l’extraction des structures détachées de notre corpus où les mots sont annotés en parties du discours. Comme les outils d’annotation morphosyntaxique peuvent être source d’erreurs, une classification manuelle s’avère nécessaire après l’extraction automatique pour assurer qu’on a saisi les exemples convenables.[4]

Le participe passé français peut avoir deux valeurs, verbale et adjectivale. On parle de la valeur adjectivale quand le participe passé fonctionne comme un attribut du sujet (par exemple Tania est surprise.) ou une épithète (par exemple une fille surprise), et de la valeur verbale quand il « exprime un procès au passif avec un complément d’agent (Ma voiture a été révisée par le garagiste) ou quand il constitue la forme composée d’un verbe actif (Il est arrivée en retard) » (Riegel, Pellat et Rioul 1999 : 344).

Par conséquent, dans notre recherche, nous classifions les participes passés français constituant le noyau d’une construction détachée dans trois groupes selon leurs fonctions :

a) le participe passé : dans la voix active, il exprime une action accomplie (d’habitude l’antériorité comme dans (5) et (6)) et a donc une fonction verbale. Ce groupe réunit les exemples de la forme composée du participe passé (exemple (5)) et ceux de la forme simple qui constituent les temps composés avec l’auxiliaire être (par exemple accouru, arrivé, devenu, resté). Les constructions détachées contenant l’une ou l’autre forme correspondent à une proposition subordonnée et possèdent une valeur circonstancielle, le plus souvent temporelle, comme c’est évident dans (5a) et (6a) :

(5) Ayant fini son travail, Paul est parti à la maison.

(5a) Quand Paul avait fini son travail, il est parti à la maison.

(6) Arrivée à la maison, Claire s’est tout de suite mise à cuisiner.

(6a) Quand Claire est arrivée à la maison, elle s’est tout de suite mise à cuisiner.

Quand le participe passé exprime l’antériorité, il se trouve toujours en position initiale (Rossi-Gensane 2009 : 188) et exprime une action accomplie par rapport au verbe principal (Riegel, Pellat et Rioul 1999 : 343). Les participes passés exprimant la simultanéité sont rares : dans ce cas-là, ils dénotent le plus souvent un état transitoire qui dure ou a duré au moment de l’action de la proposition principale (Havu 2002 : 5).

b) le participe passif : il exprime le résultat au passif et a une fonction verbale.[5] Dans la structure de surface, le complément d’agent peut être exprimé (introduit par par ou de) ou non[6] :

(7) Ma voiture a été révisée par le garagiste. / Révisée par le garagiste, ma voiture…

c) le participe d’état : exprime une propriété ou une caractéristique transitoire du référent (par exemple déçu, fatigué, connu), ou bien un état qui découle de l’achèvement d’une action (Riegel, Pellat et Rioul 1999 : 438-439) :

(8) Fatigué, Antoine ne va pas sortir.

Lorsque la forme passive du participe passé (par exemple couvert, engourdi) exprime une caractéristique ou l’état, la valeur adjectivale efface la valeur verbale (1999 : 343-344). Riegel, Pellat et Rioul (1999 : 343, 438) placent le participe d’état passif, qui est le résultat d’une action accomplie, dans le même groupe que le participe passif qui exprime une action (ils l’appellent le passif action). Comme le premier possède une valeur adjectivale et le deuxième une valeur verbale, nous les traiterons à part, car nous supposons que l’analyse des traductions slovènes puisse montrer les différences entre les deux types.

3. Méthodologie

3.1 Corpus de travail

Notre recherche est basée sur les exemples authentiques tirés du corpus parallèle français-slovène FraSloK (Mezeg 2010), plus précisément de sa partie littéraire, constituée de douze romans français contemporains, parus entre 1989 et 2006,[7] et de leurs traductions en slovène, publiées entre 1997 et 2008. Notre corpus de travail comprend 1 302 911 mots (dont 701 715 la partie française et 601 196 la partie slovène). Il représente, à ce jour, le seul corpus littéraire existant pour la combinaison linguistique français-slovène. Les textes ont été alignés au niveau phrastique par le logiciel ParaConc (Barlow 2001) et étiquetés morphosyntaxiquement – la partie française par l’étiqueteur TreeTagger (Schmid 1994) et la partie slovène par l’étiqueteur ToTale (Erjavec, Ignat, Pouliquen et Steinberger 2005).

3.2 Repérage (semi-)automatique des constructions détachées contenant un participe passé

Faute d’outils permettant l’étiquetage automatique au niveau de types de phrases et de fonctions grammaticales, la tâche du repérage automatique des structures complexes, telles que les constructions détachées, s’avère assez ardue et d’autant plus compliquée en raison des critères très restrictifs de ce qui constitue une construction détachée (la place dans la phrase, l’ellipse du sujet et d’une forme verbale personnelle dans la structure de surface, le participe passé pouvant être précédé d’un adverbe ou d’une conjonction, etc. En effet, comme le constatent déjà Benzitoun et Caddeo (2005 : 307), il est « difficile d’élaborer des requêtes à partir des critères définitoires de l’apposition[8] en vue d’une automatisation des recherches »,c’est pourquoi il n’est pas surprenant que la majorité des recherches sur les structures détachées se penchent sur les exemples repérés de manière manuelle.[9]

Pour les fins de cette étude, les exemples de constructions détachées initiales françaises formées autour d’une forme simple ou composée du participe passé ont été repérés automatiquement du corpus FraSloK par le logiciel ParaConc à l’aide de patrons syntaxiques, composés d’étiquettes morphosyntaxiques et d’expressions régulières. Bien que les patrons utilisés se soient révélés productifs, le processus de repérage automatique a apporté de nombreuses occurrences erronées qui se sont glissées parmi les résultats à cause d’étiquettes fautives et parce qu’elles ne constituaient pas une structure détachée, comme l’a révélé l’analyse manuelle laborieuse des exemples extraits. 

Prenons comme exemple l’un des 11 patrons utilisés :

(<w R>[A-Z]\w+(\W<w R>\w+){0,2}\W<w Gdr>\w+|<w Gdr>[A-Z]\w+)

Ce patron cherche tous les participes passés simples (étiquette Gdr) figurant au tout début de la phrase ou précédé de 0 à 2 adverbes (étiquette R), par exemple :

(9) Ralentis par la charge, ils mirent beaucoup de temps à regagner Port Radium. (LIT003-FRA)[10]

(10) Trop souvent mobilisée pour des tâches policières dans les territoires palestiniens occupés, l'armée n'a pas été préparée à la guerre, […]. (LMD015-FRA)[11]

Il s'est révélé le plus productif de tous les patrons, apportant 72 % (520 exemples) de toutes les constructions détachées ayant comme noyau un participe passé. Les occurrences erronées s’élevaient à moins de 30 % et étaient dues à des raisons déjà indiquées plus haut : étiquettes fautives (par exemple vu (exemple 11) et (bien) entendu étiquetés comme participes passés), des noms dont la forme égale celle d’un participe passé, étiquetés comme participes passés (par exemple passé (exemple (12)), député, parti)), et lorsque le participe passé ne figure pas dans une structure détachée (comme dans (13) où il est employé dans une phrase simple à l’impératif) :

(11) *Vu l'apparence de cette femme, c’est ce qui surprend le plus. (LIT003-FRA)

(12) *Passé et passif des universités américaines (LMD046-FRA)[12]

(13) As-tu vraiment besoin de moi ? *Réfléchis bien ! (LIT012-FRA)    

Après le tri manuel, il nous est resté 723 exemples de constructions détachées contenant un participe passé,[13] extraits de la partie littéraire du corpus FraSloK. Les exemples de la forme composée du participe passé sont très peu nombreux (26 occurrences ou 3,6 % contre 697 occurrences ou 96,4 % de la forme simple du participe passé).

3.3 Emploi des constructions détachées contenant un participe passé dans les romans français du corpus

Si l’on prend en considération le corpus littéraire dans sa totalité, il apparaît en moyenne 1 structure détachée sur 1000 mots. Or, nous voulions aussi vérifier quelle est la fréquence des constructions détachées contenant un participe passé dans des romans français individuels de notre corpus. Comme le montre le tableau 1, la répartition n’est pas équilibrée. En effet, 40 % de toutes les structures détachées sont utilisées dans le roman Impératrice de Shan Sa, où l’on trouve environ 5 constructions détachées sur 2000 mots. Deux autres romans ressortent encore selon la fréquence, notamment Le ventre de l'Atlantique de Diome Fatou qui contient 14 % de toutes les structures détachées ou bien 3,3 constructions détachées sur 2000 mots, et Je m'en vais de Jean Echenoz (10 % de toutes les constructions détachées, 2,9 constructions détachées sur 2000 mots). Pour ce qui est la durée d'un texte, le roman Un sécret de Philippe Grimbert ressort encore de la moyenne avec 3,6 constructions détachées sur 2000 mots. Les autres romans contiennent moins de 10 % de toutes les constructions détachées. En plus, la fréquence de constructions détachées sur 2000 mots est plus basse de la moyenne, ce qui nous amène à conclure que dans les textes littéraires français contemporains, l'emploi d'une structure détachée comprenant un participe passé dépend significativement du style de l'auteur.


No de CD[14]

Taux par rapport au nombre total de CD

Nombre de CD par 2000 mots

Fou de Vincent (Hervé Guibert, 1989, Éditions de Minuit) – LIT001


  0,4 %


Un secret (Philippe Grimbert, 2004, Éditions Grasset & Fasquelle) – LIT002


  6,5 %


Je m'en vais (Jean Echenoz, 1999, Éditions de Minuit) – LIT003


  10,0 %


Balzac et la petite tailleuse chinoise (Dai Sijie, 2000, Éditions Gallimard) – LIT004


 5,0 %


Truismes (Marie Darrieussecq, 1997, P.O.L. Éditeur) – LIT005


  0,1 %


Eldorado (Laurent Gaudé, 2006, Actes Sud) – LIT006


  0,3 %


Mammifères (Pierre Mérot, 2003, Éditions Flammarion) – LIT007


  1,7 %


Le ventre de l'Atlantique (Fatou Diome, 2003, Éditions Anne Carrière) – LIT008


  14,1 %


Plateforme (Michel Houellebecq, 2001, Éditions Flammarion) – LIT009


  4,6 %


Le testament français (Andreï Makine, 1995, Éditions Mercure de France) – LIT010


  9,1 %


Impératrice (Shan Sa, 2003, Éditions Albin Michel S.A.) – LIT011


  39,7 %


L'Amour du prochain (Pascal Bruckner, 2004, Éditions Grasset & Fasquelle) – LIT012


  8,6 %




100 %


Tableau 1 : Répartition des constructions détachées contenant un participe passé dans des romans français individuels du corpus FraSloK.

4. Analyse syntaxique et sémantique des traductions slovènes

Après le repérage semi-automatique, les participes passés ont été classés dans l’une des trois catégories selon leur fonction ou valeur. Comme le montre le graphique 1, les participes d’état sont les plus représentés, figurant dans plus de la moitié des exemples extraits (410 exemples ou 56,7 %) ; les participes passifs sont utilisés dans presque un tiers des structures détachées (231 exemples ou 32 %), tandis que les participes passés qui constituent la forme composée du participe passé ou forment les temps composés avec l’auxiliaire être sont les moins fréquents (82 exemples ou 11,3 %).


Graphique 1 : Fréquence des différents types de participes passés dans les constructions détachées françaises du corpus littéraire FraSloK.

Toutes les traductions slovènes ont ensuite été analysées, d’un côté, syntaxiquement pour voir par quelles structures grammaticales les traducteurs ont exprimé les structures détachées comprenant l’un des trois types de participes passés, et, de l’autre côté, sémantiquement pour vérifier si en slovène, le rapport logique entre le contenu de la structure détachée et le reste de la phrase reste implicite comme dans la langue source ou bien s’il est précisé.

4.1 Analyse syntaxique

La graphique 2, qui résume l’analyse syntaxique des traductions slovènes, montre des différences assez notables entre les trois catégories.


Graphique 2 : Répartition d’équivalents syntaxiques slovènes des constructions détachées françaises selon le type de participe passé.

Pour ce qui est la catégorie ‘participe passé’, la grande majorité (61 exemples de 82 ou 74,4 %) est exprimée en slovène par une forme verbale personnelle, utilisée dans une proposition subordonnée, ce qui n’est pas surprenant car à l’actif, il exprime en français une action verbale. Dans la majorité des traductions, son contenu est exprimé par le participe en -l (celui-ci correspond à cette catégorie de ‘participe passé’ français) qui sert à former le temps passé,[15] prédominant dans notre corpus dans les traductions de ce type de participe passé :

(14) Arrivé à ma hauteur, son porteur voulut faire une halte, et posa la chaise sur un rocher carré. (LIT004-FRA)

(14a) Ko je nosač prišel do mene, si je hotel malo oddahniti in je postavil stol na ploščato skalo. (LIT004-SLV) [Quand le porteur est arrivé à ma hauteur, il voulut…]

Nous observons le même pour la deuxième plus fréquente stratégie où le contenu du participe passé est exprimé en slovène par une forme verbale personnelle dans une phrase coordonnée (12 exemples ou 14,6 %). Les autres relations phrastiques et fonctions syntaxiques sont quasi inexistantes.

Quant aux constructions détachées contenant un participe passé au passif suivi d’un complément d’agent, le participe passif lui aussi est le plus souvent exprimé en slovène par une forme verbale personnelle au passé (le participe en -l) dans une proposition subordonnée (77 exemples de 231 ou 33,3 %), plus rarement dans une phrase coordonnée (29 exemples ou 12,5 %, exemple (15a)), à la voix active :

(15) Harcelée par leur solitude, j’étais davantage obsédée par l’absence de ce frère. (LIT011-FRA)

(15a) Bolj ko so me nadlegovale s svojimi zahtevami, bolj me je obsedala prijateljeva odsotnost.(LIT011-SLV) [Plus elles me harcelaient avec leurs demandes, plus j’étais obsédé…]

Dans quasi un tiers des traductions (catégorie ‘fonctions syntaxiques’, 63 exemples ou 27,3 %), le contenu de participes passifs est rendu en slovène soit par un groupe nominal prépositionnel ayant la fonction d’un complément circonstanciel (surtout de manière, comme dans (16a)), soit par un participe en -n ou en -t ayant la fonction d’un attribut de l’objet (exemple (17a)) ou bien par un participe en -n et en (plus rarement par un adjectif) fonctionnant comme une épithète (cette fonction est d’ailleurs moins présente que la fonction de l’attribut de l’objet). Lors des traductions slovènes, comme c’est le cas dans (17a), nous notons qu’il est souvent difficile de distinguer la passivité de l’état après une action accomplie (les deux peuvent être exprimés par le participe en -n ou en -t) :

(16) Maintenant, incité par des hommes ambitieux, ton frère la réclame. (LIT011-FRA)

(16a) Zdaj jo pod vplivom častihlepnežev zahteva tvoj brat. (LIT011-SLV) [littéralement : *Maintenant la sous l’influence des hommes ambitieux réclame ton frère. / dans l’ordre de mots français correct : Maintenant ton frère la réclame sous l’influence des hommes ambitieux.]

(17) Affecté par cette série de condamnations, il paraissait encore plus désarmé. (LIT011-FRA)

(17a) Prizadet zaradi vrste obsodb se je zdel še bolj nebogljen. (LIT011-SLV) [Affecté par cette série de condamnations il paraissait…]

Reste à mentionner encore deux stratégies qui couvrent un peu plus de 10 % d’exemples de participes passifs : la rétention de la structure détachée (le plus souvent participiale) en slovène (29 exemples ou 12,5 %) et l’emploi d’une forme verbale personnelle dans une phrase juxtaposée (exemple (18a)) ou, rarement, une phrase simple :

(18) Trompé par une fioriture inattendue de l’orchestre, je fis une fausse manœuvre et ma poitrine s’aplatit contre la sienne. (LIT010-FRA)

(18a) Zmotili so me nepričakovani dodatni takti orkestra, naredil sem napačen korak in moje prsi so se združile z njenimi. (LIT010-SLV) [Des mesures supplémentaires inattendues de l’orchestre m’ont trompé, je fis un faux pas et…]

Le plus représenté des trois types de participes passés traités, le contenu du participe d’état est dans plus d’un tiers des traductions slovènes (163 exemples de 410 ou 39,7 %) exprimé par un participe d’état en -l, -n ou -t ayant en majorité la fonction d’un attribut de l’objet[16] (60 %, exemple (19a)), moins fréquemment d’une épithète, le plus souvent antéposée au nom (20 %, exemple (20a)), tandis que les autres fonctions syntaxiques (complément circonstanciel, sujet, objet) sont rares :

(19)Penchés par-dessus la rampe, nous écarquillions les yeux en essayant de voir le plus de ciel possible. (LIT010-FRA)

(19a)Nagnjena nad ograjo sva napenjala oči, da bi videla čim več neba. (LIT010-SLV)

(20) Embarrassé, l’Empereur tapa le sol du pied et regarda autour de lui. (LIT011-FRA)

(20a) Osramočeni cesar je nekajkrat z nogo potrkal ob tla in se ozrl okoli sebe. (LIT011-SLV) [L’Empereur embarrassé tapa le sol…]

Comme s’est évident à partir du graphique 2 ci-dessus, les structures détachées sont le plus retenues dans le cas des participes d’état où c’est la deuxième stratégie la plus fréquente (110 exemples ou 26,8 % ; cela représente 78,6 % de toutes les structures détachées retenues ayant comme noyau l’un de nos trois types de participe passé). En majorité de nature participiale (exprimés par un participe en -n, -t, -l, moins souvent en ), la moitié des participes détachés slovènes gardent leur position périphérique au début de la phrase (exemple (21a)), l’autre moitié étant placée après le groupe sujet :

(21) Absorbé dans ses pensées, il s’était arrêté machinalement sans saluer. (LIT008-FRA)

(21a) Zatopljen v svoje misli, se je bil avtomatično ustavil brez pozdrava. (LIT008-SLV)

D’après l’analyse, trois traductions slovènes se détachent par rapport au nombre de constructions détachées retenues : la traduction du roman Impératrice (en slovène Cesarica, 37 % de toutes les structures détachées), L’Amour du prochain (en slovène Ljubezen do bližnjega, 21 %) et Le ventre de l'Atlantique (en slovène Trebuh Atlantika, 14 %). Il est intéressant de noter que plus de la moitié des structures détachées retenues gardent la place initiale dans la phrase. Dans les autres traductions, la rétention de ces structures est rare.

Dans seuls environ 10 % d’exemples, le contenu des participes d’état est exprimé en slovène par une forme verbale personnelle (le participe en -l) ou un attribut du sujet (souvent le participe d’état en -n ou en -t) dans une phrase subordonnée (41 exemples ou 10 %), coordonnée (51 exemples ou 12,4 %) ou autre (juxtaposée, simple ou complexe ; 36 exemples ou 8,8 %). Lorsqu’il est exprimé par une forme verbale personnelle, l’état du texte source devient en slovène une action (accomplie, comme dans (22a)) :

(22) Guérie, je ne pus plus me déplacer sans assistance. (LIT011-FRA)

(22a) Ko sem ozdravela, nisem mogla več hoditi brez pomoči. (LIT011-SLV) [Lorsque j'ai guéri, je ne pus plus…]

Pour résumer, mis à part les catégories ‘construction détachée’ et ‘omission’ (celle-ci concerne très peu d’exemples), l’explicitation syntaxique est bien évidente dans les trois catégories : elle s’élève à 98 % dans le cas des participes passés exprimant l’action à l’actif, à 84 % dans le cas des participes passifs et à 71 % dans le cas des participes d’état.

4.2 Analyse sémantique

Comparé avec les exemples sources, dans la majorité des traductions des constructions détachées contenant l’un des trois types de participes passés, son contenu est exprimé dans une phrase à forme verbale personnelle, le rapport logique qui s’instaure entre le contenu d’une structure détachée et le reste de la phrase étant explicité.

Le graphique 3 résume la nature des rapports logiques dans les traductions slovènes :


Graphique 3 : Répartition des rapports logiques exprimés dans les traductions slovènes.

Dans les traductions slovènes, l’explicitation du rapport sémantique entre le contenu d’une structure détachée française et le reste de la phrase ne peut être observée que dans le cadre des catégories ‘Subordination’, ‘Coordination’ et ‘Fonctions syntaxiques’, le rapport étant ailleurs, comme dans la langue source, implicite ou non précisé. L’interprétation sémantique des constructions détachées dépendant de nombreux facteurs (par exemple du contexte inter- et extraphrastique, de nos connaissances linguistiques et extralinguistiques), l’analyse et l’argumentation des décisions des traducteurs requerraient une discussion sérieuse dépassant le cadre de cet article, c’est pourquoi nous ne proposons qu’un aperçu rapide de la nature des rapports logiques précisés dans les traductions slovènes du corpus littéraire FraSloK.

Premièrement, d’après l’analyse, les participes passés qui constituent la forme composée ou bien simple avec l’auxiliaire être à l’actif se différencient de deux autres catégories traitées, traduisant en majorité la valeur circonstancielle par la forme verbale personnelle dans une subordonnée (65,9 %) : dans 70 % d’exemples, elle est temporelle, exprimée par le connecteur ko (quand / lorsque, exemple (23a)), suit la valeur causale avec 14,8 %, tandis que les autres valeurs circonstancielles ne sont quasi pas présentes. En deuxième place vient la valeur descriptive ou explicative qui apporte une information supplémentaire sur le référent ou exprime son état ou caractéristique en relation avec l’action principale. Elle s’instaure majoritairement quand il n’est pas possible d’établir une relation sémantique spécifique entre une construction détachée et le reste de la phrase (Combettes 2005 : 34). Cette valeur se traduit d’habitude dans notre corpus littéraire par une proposition subordonnée relative introduite par ki (qui). Nous constatons que des fois, la valeur explicative est exprimée dans certaines traductions slovènes, alors que le contexte de l’exemple source permet aussi une autre interprétation, par exemple causale, comme dans l’exemple (24)-(24b). En effet, en français, plusieurs interprétations sont souvent possibles, tandis qu’en slovène, lorsqu’une structure détachée n’est pas retenue dans le texte cible et qu’il n’y ait pas assez d’éléments dans le texte source pour déterminer quelle valeur l’emporte, le traducteur doit choisir une valeur, c’est pourquoi l’explicitation sémantique n’est fréquemment qu’une parmi toutes les interprétations possibles et relève du choix subjectif d’un traducteur. Pour conclure, l’explicitation sémantique prédomine (93 %), le taux des exemples implicites étant minime (7 %). 

(23) Arrivé à Québec, Ferrer prit un taxi de marque Subaru en direction du port, département des garde-côtiers, môle 11. (LIT003-FRA)

(23a) Ko je Ferrer prišel v Québec, se je s taksijem znamke Subaru odpeljal proti pristanišču, območje obalne straže, pomol 11. (LIT003-SLV) [Lorsque Ferrer arriva à Québec, il prit…]

(24) Ayant déjà eu recours deux ou trois fois à ses services, Ferrer le connaissait un peu, cet expert. (LIT003-FRA)

(24a) Ferrer, ki ga je nekajkrat že prosil za usluge, ga je že malo poznal, tega izvedenca. (LIT003-SLV) [Ferrer, qui avait déjà demandé quelquefois ses services, le connaissait…]

(24b)[17] Ker ga je Ferrer že nekajkrat prosil za usluge, ga je že malo poznal … [Comme Ferrer avait déjà demandé quelquefois ses services, il le connaissait…]

Deuxièmement, le corpus de participes passifs, quant à lui, témoigne d’une répartition assez homogène des rapports dans les traductions slovènes. En effet, le taux des exemples ayant une valeur explicative (37,7 %) ou circonstancielle (35,5 %) est presque le même, tandis que la fréquence des exemples gardant le rapport implicite comme dans le texte source est légèrement inférieure (26,8 %). Parmi les valeurs circonstancielles que traduisent surtout les formes verbales personnelles et les compléments circonstanciels, la valeur de temps l’emporte doucement sur les valeurs de manière (exemple (25)) et de cause.

(25) Suivie par une dizaine de servantes et de femmes d’atour, elle s’effaça entre les arbres. (LIT011-FRA)

(25a) V spremstvu deseterice služabnic in spletičen je izginila med drevesi. (LIT011-SLV) [En compagnie d’une dizaine de servantes et de femmes d’atour elle disparut entre les arbres.]

La valeur explicative, quant à elle, est portée par les attributs de l’objet et les épithètes qui qualifient le sujet, ainsi que par les formes verbales personnelles dans une phrase subordonnée ou, moins fréquemment, coordonnée (exemple (26a)). En fin de compte, attestée dans 73 % des traductions des participes passifs, l’explicitation sémantique l’emporte sur l’implicitation.

(26) Unis par la détresse, ils étaient devenus inséparables. (LIT011-FRA)

(26a) Huda žalost ju je zbližala in postala sta nerazdružljiva. (LIT011-SLV) [La détresse les avait unis et ils étaient devenus inséparables.]

Enfin, dans les traductions slovènes des constructions détachées comprenant un participe d’état prédomine la valeur explicative (45,9 %) sous forme d’un participe ayant la fonction d’un attribut de l’objet ou d’une épithète (exemple (27a)) ou, plus rarement, d’une forme verbale personnelle dans une phrase subordonnée ou coordonnée.

(27) Surpris, les moines se rendirent sans hésiter. (LIT011-FRA)

(27a) Presenečeni menihi so se brez odlašanja vdali. (LIT011-SLV) [Les moines surpris se rendirent sans hésiter.]

Dans plus d’un tiers des exemples (37,8 %), on constate que le rapport logique reste opaque comme dans la langue source, surtout en raison du grand nombre de constructions détachées retenues. Néanmoins, l’explicitation sémantique prévaut, s’élevant à 62,2 %.

Contrairement aux autres types de participes passés traités, la valeur circonstancielle est le moins attestée dans les traductions des participes d’état (16,3 %). Le participe d’état français est, dans ces exemples, d’habitude rendu en slovène par une forme verbale personnelle dans une phrase subordonnée ou coordonnée, ce qui signifie que l’état français se transforme en slovène en une action, comme le montre l’exemple suivant :

(28) À cet instant, coincé au milieu du passage, je me demandai ce que dirait le vieux Jean-Christophe, si je faisais volte-face. (LIT004-FRA)

(28a) Ko sem takole čepel sredi grebena, sem se vprašal, kaj bi rekel Jean-Christophe, če bi odnehal. (LIT004-SLV) [Lorsque je me tenais accroupi au milieu du passage, je me demandai…]

La partie slovène portant le contenu de la construction détachée française entretient le plus souvent avec le reste de la phrase un rapport de temps (exemple (28a))[18] ou de manière, des fois aussi de lieu et de cause, les autres rapports étant très rares.

5. Remarques conclusives

Le présent article a apporté une étude contrastive comparant l’emploi des structures détachées formées autour d’un participe passé dans des romans français contemporains et leurs traductions en slovène. Nous avons examiné uniquement les structures figurant au tout début de la phrase qui, vu leurs caractéristiques syntaxiques particulières et leur opacité sémantique dans la structure de surface, semblent être le plus problématiques pour un locuteur non francophone.  

Notre décision initiale de distinguer entre les trois types de participes passés, issue de la supposition qu’ils puissent, en raison de leur nature et leurs fonctions / valeurs différentes, entraîner des différences importantes dans les traductions slovènes, s’est avérée judicieuse. Globalement, l’analyse a montré l’explicitation syntaxique et sémantique dans la majorité des traductions de trois types de participe passé, les deux étant le plus élevées dans le cas de participes passés désignant une action à l’actif (98 % et 93 % respectivement), suivent les participes passifs (84 % et 73 % respectivement) et les participes d’état (71 % et 62 % respectivement).

Désignant une action à l’actif, les participes passés français gardent majoritairement leur fonction verbale en slovène, véhiculée pour la plupart par une forme verbale personnelle dans une proposition subordonnée de temps ou de cause. Quant aux participes passifs, la moitié est exprimée en slovène par une forme verbale personnelle à l’actif, dans une phrase subordonnée ou, moins rarement, coordonnée, et garde ainsi sa valeur verbale, le participe passif n’étant retenu que dans un cinquième des traductions slovènes, ce qui nous amène à conclure que le passif est plus fréquent en français qu’en slovène. En ce qui concerne les rapports sémantiques, nous constatons une homogénéité importante des trois catégories : la valeur explicative que portent les participes jouant le rôle d’un adjectif qualificatif (37,7 %), la valeur circonstancielle (le plus souvent de temps, de manière et de cause) portée par les prédicats dans des phrases actives (35,5 %), et l’opacité sémantique suite aux structures détachées gardées en slovène et les relations phrastiques non explicitées (26,8 %). Enfin, les participes d’état sont maintenus dans une moitié des traductions slovènes (par exemple sous forme d’un attribut de l’objet, une épithète), gardant ainsi leur valeur adjectivale, tandis que dans un cinquième d’exemples, ils sont exprimés par une forme verbale personnelle, signifiant que l’état du texte source devient une action en slovène. Sémantiquement, prédomine la valeur explicative ou descriptive portée par les participes retenus (45,9 %), concurrencée par l’implicitation suite au grand nombre de structures détachées maintenues en slovène (37,8 %). Non surprenant vu le petit nombre d’exemples témoignant d’une valeur verbale, la valeur circonstancielle (traduisant le temps, la manière, le lieu et la cause) est le moins attestée (aussi si on prend en considération les trois types de participes passés).

L’extraction semi-automatique nous a apporté 723 exemples de constructions détachées dans 12 romans contemporains, celles contenant un participe d’état (56,7 %) l’emportant sur celles comprenant un participe passif (32,0 %) et un participe passé ‘actif’ (11,3 %). Dans l’ensemble, la structure détachée est retenue dans moins de 20 % de traductions slovènes, en majorité dans le cas de participes d’état (78,6 % de toutes les structures détachées extraites). Selon l’analyse, l’emploi de ces structures ne ressort que dans trois romans français (surtout dans Impératrice qui en contient 40 %), ce qui nous conduit à constater que dans le registre littéraire français, l’emploi des constructions détachées dépend du style de l’auteur. Par conséquent, le nombre de structures retenues dans une traduction peut aussi dépendre de la quantité de telles structures dans le texte source, ce qui est le cas dans notre corpus où les traductions de mêmes trois romans français se détachent, conservant ainsi le style du texte source. D’ailleurs, il serait nécessaire d’examiner dans le futur l’emploi des structures détachées dans les textes slovènes authentiques pour évaluer l’influence de la langue source sur la langue de traduction. 

La méthode de l’extraction informatique des constructions détachées du corpus s’est avérée laborieuse, d’un côté en raison du processus complexe de l’élaboration des patrons syntaxiques, c’est pourquoi elle est plus adéquate pour la recherche des mots individuels que des structures complexes. De l’autre côté, lors de l’analyse manuelle, nous avons fait face au problème d’étiquetage automatique des mots français suite aux étiquettes fautives attribuées à certains ‘participes passés’, présentant ainsi de nouveaux défis pour les chercheurs du domaine de traitement automatique du langage naturel. Cela mis à part, le corpus littéraire utilisé, d’autant plus grâce à son étiquetage en parties du discours, représente une source riche de textes authentiques français et de leurs traductions en slovène, permettant ainsi un tas de recherches différentes et étant donc plus utile et durable des listes d’exemples d’un certain phénomène linguistique, dressées à la main.


Bajec, Anton, Rudolf Kolarič et Mirko Rupel (1964). Slovenska slovnica (Druga popravljena izdaja). Ljubljana: Državna založba Slovenije.

Barlow, Michael (2005). ParaConc. Houston: Athelstan.

Benzitoun, Christophe, et Sandrine Caddeo (2005). « Indices linguistiques pour le repérage automatique de certains types d’apposition dans un corpus de presse ». Dans Williams, Geoffrey (éd.) : La linguistique de corpus. Rennes : Presses Universitaires de Rennes. 307-321.

Boch, Françoise, Agnès Tutin et Damien Laurent (2009).« Construction détachée et adjectifs d’affects ». Dans Apothéloz, Denis, Bernard Combettes et Franck Neveu (éd.) : Les linguistiques du détachement. Actes du colloque international de Nancy, 7 à 9 juin 2006. Berne : Peter Lang. 99-115.

Breznik, Anton (1967). Življenje besed (adapté par Jakob Šolar). Maribor : Obzorja.

Combettes, Bernard (1998). Les constructions détachées en français. Paris : Ophrys.

Combettes, Bernard (2005). « Les constructions détachées comme cadres de discours ». Dans Langue française 148 (Les adverbiaux cadratifs). 31-44.

Erjavec, Tomaž, Camelia Ignat, Bruno Pouliquen et Ralf Steinberger (2005). « Massive multi-lingual corpus compilation: Acquis Communautaire and totale ». Dans Vetulani, Zygmut (éd.) : Proceedings of the 2nd Language and Technology Conference, Poznan, 21 à 23 avril. Poznan: Wydawnictwo Poznanskie Sp. z.o.o. 32-36.

Forsgren, Mats (1993). « L'adjectif et la fonction d'apposition : observations syntaxiques, sémantiques et pragmatiques ». Dans L'Information grammaticale 58. 15-22.

Forsgren, Mats (2000). « Apposition, attribut, épithète : même combat prédicatif ? ». Dans Langue française 125 (Nouvelles recherches sur l'apposition). 30-45.

Havu, Eva (2002). « L’interprétation des constructions détachées ». Dans Cercles de linguistique appliquée à la communication 10. http://www.ucm.es/info/circulo/no10/havu.htm (Consulté en août 2016).

Jackiewicz, Agata, Thierry Charnois et Stéphane Ferrari (2009). « Jugements d’évaluation et constituants périphériques ». Dans Actes de la 16ème Conférence sur le Traitement Automatique des Langues Naturelles, Senlis, 24 à 26 juin 2009. Actes électroniques. 1-10.

Le Nouveau Petit Robert de la langue française (2008). Paris : Le Robert.

Mezeg, Adriana (2010). « Compiling and Using a French-Slovenian Parallel Corpus. Dans Xiao, Richard (éd.) : Proceedings of The International Symposium on Using Corpora in Contrastive and Translation Studies (UCCTS 2010). Ormskirk: Edge Hill University. http://www.lancs.ac.uk/fass/projects/corpus/UCCTS2010Proceedings/papers/Mezeg.pdf (Consulté en août 2016).

Mezeg, Adriana (2011). Korpusno podprta analiza francoskih polstavkov in njihovih prevedkov v slovenščini. Thèse de doctorat, Université de Ljubljana, Faculté des Lettres, Département des langues et littératures romanes.

Milojević Sheppard, Milena (1991). Oblikoslovno-skladenjske razširitve pri prevajanju iz angleščine v slovenščino kot prototipični odziv na zapletenost izvirnika. Thèse de doctorat. Ljubljana : Facultés des Lettres.

Neveu, Franck (1996). « La notion d'apposition en linguistique française : perspective historique ». Dans Le français moderne 64 : 1. 1-27.

Riegel, Martin, Jean-Christophe Pellat et René Rioul (1999). Grammaire méthodique du français. Paris : Presses Universitaires de France.

Rossi-Gensane, Nathalie (2009). « Register variation in the non-standard use of non-finite forms ». Dans Beeching, Kate, Nigel Armstrong et Françoise Gadet (éd.) : Sociolinguistic variation in contemporary French. Amsterdam, Philadelphie : John Benjamins Publishing Company. 177-191.

Schmid, Helmut (1994).« Probabilistic Part-of-Speech Tagging Using Decision Trees ». Dans : Proceedings of the International Conference on New Methods in Language Processing, Manchester, Grande Bretagne. http://www.cis.unimuenchen.de/~schmid/tools/TreeTagger/data/tree-tagger1.pdf (Consulté en août 2016).

Toporišič, Jože (2004). Slovenska slovnica (Četrta, prenovljena in razširjena izdaja). Maribor : Obzorja.

Trésor de la langue française. http://atilf.atilf.fr/tlf.htm (Consulté en août 2016).


[1] A part une forme verbale, la prédication peut aussi se former autour d’un groupe nominal ou d’un adjectif (cf. Combettes 1998).

[2] Il n'existe pas encore, à notre connaissance, d’études portant sur l’emploi des structures détachées dans des textes slovènes authentiques contemporains.

[3] Breznik ne s'appuie pas sur des recherches concrètes mais plutôt sur son « sens de la langue ». Les autres grammaires prescriptives (par exemple de Bajec, Kolarič et Rupel 1964 : 211, 261-262, ou bien de Toporišič 2004 : 632) signalent qu’une construction détachée peut être transformée en une phrase (par exemple simple, subordonnée, coordonnée) comportant une forme verbale personnelle, mais aucune ne détermine que la phrase slovène doit obligatoirement contenir un verbe.

[4] Certaines formes peuvent, entre autres, correspondre à un participe passé et à un adjectif. Prenons comme exemple les formes déçu (du verbe décevoir) et né (du verbe naître) : dans le dictionnaire informatisé Trésor de la langue française, on trouve les formes participiale et adjectivale sous l’entrée du verbe (décevoir et naître respectivement), tandis que dans Le Nouveau Petit Robert (2008), les formes adjectivales ont une entrée indépendante (déçu et  respectivement). Les deux formes (c’est-à-dire déçu et ) sont convenables pour notre recherche en raison de leur origine verbale.    

[5] Il a la même forme que le participe d’état, mais vu sa fonction différente, nous le traitons à part.

[6] On parle du passif incomplet (Riegel, Pellat et Rioul 1999 : 439). Dans grand nombre de phrases passives en français, le complément d’agent n’est pas exprimé (ibid.).

[7] Pour la liste des romans français inclus, voir le tableau 1 ci-après.

[8] Certains chercheurs appellent une construction détachée (Combettes 1998) une apposition (par exemple Neveu 1996, Forsgren 1993, 2000). Au passé, il y a eu de nombreux débats sur la dénomination de ces structures (cf., par exemple, Neveu 1996, Forsgren 1993, 2000). Nous préférons utiliser le terme de Combettes, c’est-à-dire la construction détachée, qui définit et délimite clairement la structure grammaticale et dont l’usage semble assez fréquent en linguistique française (cf. Havu 2002, Boch, Tutin et Laurent 2009, Jackiewicz, Charnois et Ferrari 2009).

[9] Le repérage automatique de certains types de constructions détachées est pourtant attesté chez Jackiewicz, Charnois et Ferrari 2009, et chez Boch, Tutin et Laurent 2009.

[10] Le code entre crochets apporte l'information d'où est pris un certain exemple. LIT003-FRA correspond au roman français numéroté de 3 (cf. Mezeg 2011, Annexe 1).

[11] Cet exemple est tiré de la partie journalistique du corpus FraSloK et n’est mentionné que pour illustrer le patron syntaxique exposé.

[12] Il s'agit d'un titre du Monde diplomatique.

[13] A cette phase, nous parlons uniquement de la forme du participe passé et ne différencions pas encore les trois types de participe passé selon leurs valeurs (cf. 4 Analyse syntaxique et sémantique des traductions slovènes).

[14] ‘CD’ signifie ‘construction détachée’.

[15] Le temps de la subordonnée dépend aussi du verbe de la proposition principale : quand celui-ci est au présent, le verbe de la subordonnée en slovène est aussi, le plus souvent, au présent : Arrivé à Toulouse en pleine nuit, Baumgartner dépose la fille devant la gare… (LIT003-FRA) En slovène : Ko Baumgartner sredi noči prispe v Toulouse, odloži dekle pred železniško postajo.(LIT003-SLV).

[16] Le plus souvent en position initiale, avant le verbe principal. Dans ce cas-là, seule une virgule non exprimée après l’attribut de l’objet différencie celui-ci d’une structure détachée. Par exemple, si dans (19a), il y avait une virgule après l’attribut de l’objet Nagnjena nad ograjo, on parlerait d’une structure détachée. En évitant la virgule, la simultanéité de l’état et de l’action principale est plus soulignée.

[17] Une interprétation possible.

[18] Dans cet exemple, le complément circonstanciel À cet instant aide à déterminer le rapport temporel.  

About the author(s)

After her studies in translation (Slovenian-English-French), Adriana Mezeg was a junior researcher at the Department of Translation Studies, Faculty of Arts, University of Ljubljana. In 2011, she defended her PhD thesis in the framework of which she compiled a 2.5-million parallel corpus FraSloK, the first corpus of its kind for the French-Slovenian language pair. Since 2007, she has been working at the Ljubljana Department of Translation Studies as a teaching assistant for the French language. Her research work focuses on contrastive linguistics and translation studies using a corpus-based approach. She is co-author of two university manuals for the French grammar, La morphosyntaxe et la sémantique du verbe français and La syntaxe du français – approche contrastive.

Email: [please login or register to view author's email address]

©inTRAlinea & Adriana Mezeg (2017).
"Participe passé détaché dans des romans contemporains français et leurs traductions slovènes"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2256

A corpus-bases study of The Translation of Point of View in four Chinese Translations of Hemingway’s The Old Man and the Sea

By Elaine Y.L. Ng (Southern University of Science and Technology, China)

Abstract & Keywords

The paper applies the corpus methodology to examine an extract from Hemingway’s (1952) The Old Man and the Sea and its four Chinese translations from the perspective of modality. Simpson’s (1993) model of point of view built on the basis of a modal grammar is employed as the linguistic tool for textual analysis. Specifically, this research compares a text from the third day’s battle of the old man with the marlin and its corresponding four translations with regard to a range of the deontic, epistemic, and boulomaic modal operators as well as the various modes of point of view. The corpus-based study aims to classify translation shifts and capture the stylistic tendencies of the four translators in their translation of the selected literary text studied. The analysis finds noticeable translation shifts and different stylistic traits among the four versions of the novella in the rendering of the selected lexical features investigated.

Keywords: Simpson’s 1993 point of view model, corpus-based translation studies, deontic, epistemic and boulomaic modals, translation of point of view

©inTRAlinea & Elaine Y.L. Ng (2017).
"A corpus-bases study of The Translation of Point of View in four Chinese Translations of Hemingway’s The Old Man and the Sea"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2255

1. Introduction

The paper aims to examine an extract from Hemingway’s (1952) The Old Man and the Sea and its four Chinese translations from the perspective of modality. The four Chinese translations of The Old Man and the Sea investigated are respectively the one produced by Hai Guan (1957), Wu Lao (1987), Li Xiyin (1987) and Zhao Shaowei (1987). Simpson’s model of point of view (1993) is employed as the linguistic tool for conducting comparative multilayered stylistic analyses. The source text will be compared to the four target texts with the aim of classifying translation shifts and capturing the four translators’ individual style. The extract mainly deals with the repeatedly unsuccessful attempts of the old man to hook the fish and his eventual success in harpooning it, yet with the tragic ending of the fish being eaten by the sharks. This part of the novella uses a variety of modal operators to represent therather paradoxical worldview of the old man in battling with the marlin and this is the reason why it was considered appropriate for stylistic study of modality in translation. This article explores how the old man encodes in language his worldview, attitudes and beliefs towards the situations around him and the way these are represented in the four translated versions of the novella.

The small parallel corpus of The Old Man and the Sea was compiled manually. The source and target fragments analysed consist of a total of 111 clauses extracted from the third day’s battle of the old man with the marlin. The text investigated consists of about 22 pages, accounting for roughly one fourth of the book examined.The analysis relies on basic corpus tools, including word counts, frequency lists, and Key Word in Context (KWIC) concordances that could be processed manually to uncover translation shifts through systematic comparison of the source text and its four versions. The quantitative findings generated from the corpus study will serve as a starting point for formulating research questions worthy of further investigation. The study presented here focuses on the following linguistic features:

  1. The deontic modals “must,” “can/ could,” “will/ would.”
  2. Epistemic modal operators including “perhaps,” “will,” “may/ maybe” “would rather,” “I do not/ cannot know,” “I wonder,” “might as well have been,” “I am not sure,” “I suppose,” “should have,” “might have,” “But I am sure,” “could have,” “But if I had,” “what he could do,” “But if I had,” “What will you,” “if they come in,” “What can you do,” “Maybe I’ll,” “What could I,” “You might,” “will probably” and “what can.”
  3. The boulomaic modal operator “wish.”
  4. Speech and thought presentation including the stands of Narrative Report of Action (NRA), Direct Speech (DS), Free Direct Speech (FDS), Direct Thought (FD), Free Direct Thought (FDT) and Indirect Thought (IT).

The paper is divided into four parts. Following on from this introduction, I will provide an overview of the core concepts of the three types of modality and the different categories of point of view based primarily on Simpson (1990, 1993). A third section presents the analyses of the source text and the target texts. The conclusion reflects on the overall patterns of choices uncovered in each of the four translations and provides suggestions for future research.

2. Deontic, Epistemic and Boulomaic Modality

Modality refers to “attitudinal” features of language and belongs to the interpersonal dimension of language. It is concerned with a speaker’s attitude towards and opinions about the events around him (Simpson1993: 47).Simpson (1993) broadly divides modality into the deontic, epistemic, and boulomaic systems. He suggests that deontic modality is the modal system of “duty,” expressing a speaker’s sense of obligation towards the accomplishment of certain tasks. Deontic modality is often realised by modal auxiliaries such as “may,” “should” and “must,” showing varying degrees of commitment attaching to the performance of certain actions, as found in the following examples given by Simpson (1990: 67) which illustrate a continuum of commitment ranging from (1) “permission” through (2) “obligation”to “requirement”:

  1. You may leave at ten o’clock.     (permission)
  2. You should leave at ten o’clock.   (obligation)
  3. You must leave at ten o’clock.     (requirement)

Epistemic modality expresses a speaker’s degree of confidence in the truth of a proposition stated. It is generally realised by modal auxiliaries such as “could,” ‘may,” “must,” “might” and “should” (Simpson1993: 48-49). Similar to deontic modality, there is an epistemic modality scale realising a semantic continuum of probability ranging from (1) uncertainty through (2) probability, (3) certainty to (4) logical necessity as illustrated by the following examples provided by Han (2005: 12):

  1. They may be there now. (possible but uncertain)
  2. They should be there now. (probable)
  3. They will be there now. (certain)
  4. They must be there now. (logical necessity)

Boulomaic modality is closely related to the deontic system. It is concerned with the wishes and desires of a speaker, as revealed in the examples below given by Simpson (1993: 48):

  1. I hope that you will leave.
  2. I wish you’d leave.
  3. I regret that you’re leaving.

The use of the modal lexical verbs “hope,” “wish” and “regret” in the above sentences all express the speaker’s wish.

It is important to notice that the boundaries between deontic and epistemic modalities are often not discrete but tend to overlap since the same modal auxiliaries may express diverse meanings that cut across both deontic and epistemic categories. The notion of the basic indeterminacy of the deontic and epistemic meanings has been observed by Coates (1983), as reviewed by Simpson (1990). Coates (1983) suggests that in the study of deontic and epistemic modalities, there are often indeterminate cases where it is hard to decide which of the two meanings is intended.

Coates provides the following example as an illustration (Coates1983: 16, quoted in Simpson1990: 68): “He must understand that we mean business.”

Coates argues that this sentence could be used to express the epistemic meaning that “Surely he understands…” or the deontic meaning that “He will have to understand….” Thus, it appears that there are cases in which both epistemic and deontic meanings are “mutually compatible” and “have undergone a merger” in a specific context (ibid.).

In addition to the overlap of the deontic and epistemic categories, Coates also points out that there is often a gradient of meaning attaching to the same modal verb; for instance, the deontic modal “must” exhibits a spectrum of meaning ranging from the “strong” or “core” meaning to the “weak” or “peripheral” meaning, as revealed in the sentence “You must eat your dinner” in which a strong sense of requirement is expressed from a mother to a child, while the sense of requirement is much weaker in the sentence “Clay pots… must have some protection from severe weather” (Coates1983: 21, quoted in Simpson1990: 68).

3. Speech and Thought Presentation

This section provides a brief review of the different categories of speech and thought presentation introduced in Simpson’s (1993) point of view model. Simpson (1987) conducted a multilayered stylistic analysis of the narrative structure of a passage of forty-four sentences from the central section of The Old Man and the Sea. He employed three different linguistic frameworks to study the grammatical structure (layer 1), textual component (layer 2) and modes of speech and thought presentation (layer 3) of the extract. With insight gained from Simpson’s (1987)stylistic analysis of an extract from The Old Man and the Sea, this study selected examples from the mini corpus of The Old Man and the Sea compiled to illustrate the different categories of speech and thought presentation, including Narrative Report of Action (NRA), Direct Speech (DS), Free Direct Speech (FDS), Direct Thought (FD), Free Direct Thought (FDT) and Indirect Thought (IT). The following briefly reviews the above-mentioned modes of speech and thought presentation.

3.1 Narrative Report of Action (NRA)

The NRA strand provides an external narrative framework around which the modes of speech and thought are woven (Simpson1993: 27). Events and actions are described by a narrator who is an outsider to them, in the third person, thus exhibiting external focalization, as in the following example:

  1. …and from the way the line slanted he could tell the fish had risen steadily while he swam (no. 1298)[1]
  2. He was past everything now and he sailed the skiff to make his home port as well and as intelligently as he could. (no. 1311)

3.2 Direct Speech (DS) and Free Direct Speech (FDS)

Direct Speech (DS) is speech reported verbatim. It is characterised by the presence of an introductory reporting clause and a reported clause enclosed in quotation marks (Simpson1993: 22); for example:

  1. ‘But man is not made for defeat,’ he said. ‘A man can be destroyed but not defeated.’ (no. 1334)
  2. ‘I could not fail myself and die on a fish like this,’ he said. (no. 1299)

A DS presentation may be stripped of its reporting clause or its quotation marks to introduce Free Direct Speech (FDS), as shown in the following example provided by Simpson (1987: 211) in his multilayered analysis of an extract from The Old Man and the Sea:

  1. “How does it go, hand? Or is it too early to know?”

There is no reporting clause in the above speech presentation, producing a freer form of DS, where it is to a greater degree liberated from the narratorial control of report. It is also possible to remove the quotation marks together with the reporting clause to produce the maximally free form of FDS presentation (Simpson1987: 208; Simpson1993: 22), such as changing the above example of FDS to – How does it go, hand? Or is it too early to know?, where the character speaks for himself with little authorial interference.

3.3 Direct Thought (DT), Free Direct Thought (FDT) and Indirect Thought (IT)

The presentation of thought is similar to the presentation of speech in terms of the categories available for thought presentation. Nevertheless, as pointed out by Leech and Short (1981: 337), the presentation of thoughts, even in the most indirect form, is just an “artifice” and a greater degree of “novelistic licence” is required to present the thoughts of a character than is needed in the presentation of speech. The following are examples of Direct Thought (DT), Free Direct Thought (FDT) and Indirect Thought (IT):

  1. ‘I should have brought a stone.’ You should have brought many things, he thought. (DT) (no. 1359-1360)
  2. I wonder what a bone spur is, he thought. (FDT) (no. 1345-1346)
  3. But if I had, and could have lashed it to an oar butt, what a weapon. (FDT) (no. 1368-1373)
  4. and he thought that now, soon, he must hit the edge of the stream (IT) (no. 1297)

Example 1 contains an unmarked reporting clause “he thought” for the first thought presentation within the same paragraph, which is enclosed in quotation marks and thus can be considered a DT. Example 2 omits the quotation marks yet retains the reporting clause, and thus can be viewed as a moderately free form of FDT. Example (3) is stripped of both the reporting clause and the quotation marks, and thus can be considered the maximally free from of FDT. This is close to the stream-of-consciousness technique, signaling an intrusion into the consciousness of the old man and the interior monologue in which he is engaged. Finally, example 4 has an introductory reporting ‘that’ clause, explicit subordination, and a declarative form for the reported clause, and thus can be classified as the IT mode of thought presentation. Since it is impossible to see inside the minds of other people, both DT and FDT modes are viewed as more “artificial”, requiring greater interpretative control than the IT mode (Leech and Short1981: 344-345).

4. Analysis of the Source Text

This section presents the analysis of the deontic, boulomaic and epistemic modalities coupled with the various modes of speech and thought presentation in The Old Man and the Sea. The fragment chosen for study abounds in deontic and epistemic modal operators showing, on the one hand, the old man’s strong sense of obligation and will power to catch the marlin and, on the other, his uncertainty and bewilderment when faced with successive defeats to hook it and subsequent attacks of the sharks to eat the whole marlin he eventually kills. I will first begin with the analysis of the deontic and boulomic modal operators.

4.1 Deontic and Boulomic Modal Operators

The modal operators “must” and “will/ would” are the most frequently used, each with a total of twenty instances, followed by “can/ could,” with a total of fourteen instances. The deontic modal expressions “I must,” “I will” and “I can/ could” all express intrinsic deontic modality, revealing the old man’s strong degree of commitment to fulfil his mission of hooking the fish . A distinction can be made between the modal auxiliaries “must” and “will,” whereby the former generally communicates a higher value of deontic modality, expressing a stronger sense of duty than the latter. The stronger deontic modal operator “must” is particularly prominent in the early phase of the third day’s battle. As he is successively defeated in his attempts to hook the fish, and starts having moments of doubt, the deontic modal auxiliary “must” becomes less common and a series of “will,” are used in expressions such as “I’ll take you” (1316), “I’ll just steer south,” (1317), “I’ll pull him,” (1318), “I’ll try it,” (1319) and “I will try it” (1320) . Therefore, while the two modal auxiliaries express the old man’s sense of duty to fulfill his mission of hooking the fish, they are used to convey different degrees of involvement at different points of the battle. The modal operator “can/ could” is used mostly to convey deontic ability, expressing the old man’s self-driven will to control the circumstances. Expressions such as “I cannot fail myself” (1299), “I can control mine,” (1300) and “I can last”(1302)are used to convey the old man’s strong will power to overcome his physical limitation and overcome the circumstances.

There is one instance – “they must have taken a quarter of him and of the best meat” – (1293) in which “must” is used not to convey deontic duty but to convey epistemic certainty, describing the old man’s assessment of the situation regarding the portion of the marlin that has been torn by the sharks. There are also three instances in which the modal auxiliary “will” – “Then in two or three turns more I will have him” (1315), “I’ll take you at the turn” (1316) and “This time I’ll pull him over” (1318) – can be interpreted as both the old man’s projected self-command to hook the fish (i.e. “will” is used in the deontic sense) and his assessments of the probability of hooking the fish (i.e. “will” is used in the epistemic sense). Moreover, there are seven instances in which “will” is used mainly to convey epistemic probability, expressing the old man’s judgment of the likelihood of the events occurring, such as the healing effect of water (1322), the coming of more sharks (1325), the coming of night time and the possibility of seeing lights (1328-1329). Finally, there are two instances of “would” that are used to express the old man’s wishes (1330, 1331).

Regarding the use of the modal auxiliary “can / could,” there are four instances in which they can be interpreted as either deontic ability or epistemic possibility. For example, the clauses “what can I think of now?” (1304), “He could only use it effectively with one hand because of the grip of the handle” (1305) and “I could have in my time” (1306) can be interpreted, respectively, as referring to the old man’s ability to think, to use the oar handle with one hand and to kill the fish in his prime time, or to the possibility of all these events occurring. Furthermore, there are four instances in which the modal auxiliaries “can/ could” are used primarily to convey epistemic possibility, expressing the old man’s assessments of the situations regarding the fish’s circling (1298), his possibility of killing the sharks by using a bat with two hands (1307-1308) and his distance from the harbour (1309). Also, the epistemic modal operators “could have” and “surely” used in the sentence – “If I could have used a bat with two hands I could have killed the first one surely” (1307-1308)– deserve attention. These are clauses governed by the logical condition that the old man could have used a bat with two hands. The use of the epistemic operators indicates a very low degree of confidence since they are expressed in the past subjunctive denoting a hypothetical rather that a real situation. The non harmonic combination of the epistemic modals “could have” with the ironic use of the modal adverb “surely” at the end of the clause serves to qualify further the statement, indicating more the old man’s tentativeness rather than his conviction towards the truth of the proposition expressed.

Finally, there are eight clauses including the boulomaic modal operator “wish,” expressing the old man’s wish that the painful reality of losing his fish did not happen. He repeats three times that he wishes the reality was just a dream and that he had never hooked the fish (1382, 1283, 1388). The old man seems to seek consolation in dreams when faced with the painful reality of losing the fish he eventually hooks.

4.2 Epistemic Modal Operators

A variety of epistemic modal devices are used throughout the corpus compiled to express the old man’s uncertainty and bewilderment in battling with the fish and the sharks. The modal auxiliaries most frequently used are “may,” “maybe,” “might have,” “might as well have been” and “should have,” which account for seventeen of the total of forty-six epistemic modal operators found in the corpus. Generally they convey epistemic possibility rather than certainty in the old man’s assessments of the likelihood of his hooking the fish and defending it from the sharks’ attacks. The use of tentative modal auxiliaries reveals the old man’s doubts regarding his control over the circumstances. In his most vulnerable moments, he thinks that the reality is just a dream, and that he can only count on luck to help him and hope that things may turn out well (1362, 1363, 1374-1375).

Other than those mostly frequently occurring modal auxiliaries mentioned-above, there are four instances of the modal auxiliary “will,” which are used to convey epistemic possibility regarding the chances of the old man seeing the fish (1336), having luck (1374), other fishermen worrying about him (1364)and being attacked by the sharks again (1379). In the modalized expression “They will probably hit me again” (1379), the modal auxiliary “will” is used in combination with the modal adverb “probably” to weaken the epistemic commitment attaching to the propositional information. Other than the use of the modal “will” to convey epistemic possibility, there is one other instance in which “will” is used to convey deontic duty ­­– “But I will try it once more” (1344), expressing the old man’s recurrent inclination to try once again to catch the fish despite his repeated failures to do so.

The epistemic possibility conveyed by the use of the modal auxiliaries as listed above is reinforced further in some cases by the use of the modal adverb “perhaps” and in others by a range of modal lexical verbs. There are five instances of the use of the modal adverb “perhaps” to indicate the old man’s partial detachment from the truth of the propositions expressed – “Perhaps in an hour I will see him” (1355), “perhaps it was a dream” (1347), “Perhaps not….Perhaps I was only better armed” (1348-1349), “and perhaps it is just a noise” (1350) –. Modal lexical verbs used include “I do not / cannot know” (1342, 1343, 1355), “I wonder” (1345, 1353), “I suppose” (1358); together with the adjectival construction “sure that” found in the modalized expressions “and I am not sure that I believe in it” (1356) and “But I am sure he would have confidence” (1364). All these epistemic modal operators tend to be pre-posed, governing the subsequent propositions stated, and thus qualifying them to express a very low degree of commitment to the truth of the propositional content. For example, in the last two sentences, the two statements “I believe in it” and “he would have confidence” are respectively qualified by the pre-posed modalized adjectival construction “I am not sure that” and “But I am sure,” detaching the speaker from making full commitment to the following propositions expressed. The old man actually is not sure about the modal assessments stated.

Finally, there is a series of other modalized expressions used in the form of interrogatives to express the old man’s self-doubt:  “But do you think…” (1354), “what he could do to a shark if he were….” (1366), “But if I had, and could have …” (1368-1369), “What will you do now…if they come in? What can you do?” (1370-1373). The use of the conditionals “if” is prominent in these clauses, and this series of expressions all use the Free Direct Thought mode to present the old man’s first-person view, highlighting his most vulnerable moments when he is on the verge of losing the battle.

4.3 Speech and Thought Presentation

There are altogether twenty-two instances of Direct Speech (DS), four of Narrative Report of Action (NRA), two of Indirect Thought (IT) and eighty-three of Free Direct Thought (FDT) in the fragment investigated. Events are mainly narrated from a position outside the consciousness of the old man via a non-participating narrator, as realized by the NRA and DS strands. The use of DS is characterized by the presence of introductory reporting clauses (usually placed in the first clause of every paragraph of speech presentation), and reported clauses enclosed in quotation marks. The reporting clauses are alternately “he said” and “the old man said.” In addition, the DS clause in clause 1384 contains some narrative report of action in its reporting clause – “the old man said after he had checked the lashing on the oar butt.” Events are mediated through the consciousness of the old man in the mode of FDT, The thoughts of the old man are presented uniformly in the form of FDT but not in the maximally free type since, even though the reported clauses are all stripped of the quotation marks, they retain the reporting clause in the first thought presentation within every paragraph of thought presentation (as is the case in reported speech). The reporting clause is formed alternately by “he thought” and “the old man thought,” except for clause1319 in which the reporting verb “thought” is changed to “promised.” There are two instances in which the reporting clauses are removed to produce the maximally free form of FDT. This occurs in clauses 1356-1358, when the old man encounters the shark’s attacks on the marlin, and clauses 1368-1373, when he laments the loss of the marlin. The use of FDT mode signals intrusion into the active mind of the old man. He is allowed to express his sequences of thoughts freely without explicit authorial intervention.

5. Analysis of the Target Texts

In this section I categorise the translation shifts identified in the four versions on the basis of Simpson’s (1993) point of view model, and provide quantitative analyses regarding the adjustments techniques used by the four translators in rendering the specific linguistic features investigated. I found prominent shifts in the four Chinese translations and obvious differences are to be found across the four translators’ approaches. Generally, the translation strategies used can be classified into four types: (1) omission, (2) addition, (3) modification and (4) restructuring of the sequences of details. Omission refers to the deletion of details, reproducing relatively reduced pictures of description. By contrast, addition refers to the incorporation of new features into the target text that do not appear in the source text. They are mostly presuppositions or inferences made explicit on the basis of the given information in the original to explicate the specific contexts in which events take place. As for modification, the original images are transposed to new ones to create different descriptions. It should be noted that the boundary between addition and modification is sometimes notclear. In some cases, a rendering can be viewed as an addition or a modification, depending on the angle adopted in the analysis. Finally, the category ‘restructuring of the sequences of details’ describes a situation where the same amount of information is basically preserved in the translation, yet its sequence is altered to produce different emphases.

5.1 Translation of the Deontic and Boulomaic Modal Operators

As shown in Table 1.2, most of the deontic and boulomaic modal operators are preserved in all four Chinese versions, though noticeable adjustments of some of the modal operators are found in the versions by Hai, Li and Zhao. Liappears to make the most adjustments in the translation of the deontic and boulomaic modal operators(twenty-three instances), followed by Zhao (fifteen instances), and Hai (twelve instances). Wu appears to be the most faithful to the original, with only one instance of adjustment made. Table 1.1. below provides a list of the translations of the deontic and boulomaic modal operators found in the four versions to show how they are generally rendered. .

must / mustn’t

一定, 必須, 得, 可不能, 該, 決不要, 就

can / could

可以, 能, 得, 必須, 該, 會, 準能, 準

will / would

要, 會, 還得, 就要, 就能, 就得, 就是, 就成, 還要


想, 但願, 真希望, 情願, 真盼望, 最好, 還不如, 倒情願, 渴望, 巴不得

Table 1.1     Chinese Translations of the Deontic and Boulomaic Modal Operators in the The Old Man and the Sea Mini Corpus

Translation strategy / frequency




















Restructuring of the sequences of details





Total frequency





Table 1.2     Translation strategies used for of Deontic and Boulomaic Modal Operators

It is noteworthy that the deontic modal “will” in clause 1315 is interpreted differently among the four translators – Hai translates it into“要” to convey the deontic meaning of duty, while Wu and Li both translate it into “就能” and Zhao into “就” to convey the epistemic meaning of probability. Likewise, the deontic modal “will” in clause 1318 is interpreted differently by Hai and Zhao – Hai translates it into “會” to convey the epistemic meaning of probability, Wu and Li simply omit it, whereas Zhao translates it into “要” to convey the deontic meaning of duty. The different interpretations of the modal “will” among the four translations could be a result of the overlap between deontic and epistemic modalities discussed above; the same modal can often be used to convey both deontic and epistemic meanings.

Wu’s translation is the most faithful among the four versions. He made only a slight adjustment by omitting the deontic modal ‘ll’ in expressing the old man’s attempt to pull the fish over to him again (1318).Next to Wu is Hai, there are altogether twelve adjustments made, ten of which are restructurings of the sequences of details, coupled with one addition and one omission. The ten restructurings of the sequences of details are all contributed by fronting the reporting clauses “he thought” and translating them either as “他想:” (a total of eight out of the ten instances) or “他想,” (a total of two out of the ten instances) at the beginning of every thought presentation. In addition, the extensive use of a colon after the reporting clause signals the presence of an omniscient narrator in the thought presentation, which is not found in the original reporting clause of the FDT presentation. Overall, Hai’s rendering of the reporting clauses of the FDT presentation is worthy of attention since it is unique among the four translations examined. As for the one instance of modification, it is a noticeable shift made by changing the reporting clause “the old man promised” in a Free Direct Thought presentation to “但他又下了決心:” (1319-1320). It highlights the old man’s strong determination to try catching the fish once again. Similarly, the recreation of the colon “:” in the presentation of the old man’s thought signals greater authorial intrusion into the old man’s consciousness by indicating the presence of the narrator as it is commonly used in reporting speech in Chinese discourse. Regarding addition, there is one instance of the conjunction “but” and the particle “ah” – “可是啊,” coupled with a repetition of the verb “撐” associated with the modal auxiliary “must” – “可是啊,你不撐也得撐” (1302) – added to emphasise the old man’s deontic duty to endure the battle with the fish.

As for Li’s translation of the deontic and boulomaic modal operators, he made the most adjustments among the four versions studied. There are a total of twenty-three instances of adjustments made, fourteen of which are modifications of the modal auxiliaries, the boulomaic modal operator “wish” and the reporting clauses, together with changing the original declarative statement to interrogative. Regarding the modification of the modal auxiliaries, two of which concern the modal auxiliaries “must” (1291, 1292); one concerns the modal auxiliary “will” – changing “But I will try it” to “我就不客氣” (1327, meaning I will not be friendly); and one is the modification of the boulomaic modal operator “wish” – changing “I wish I had a stone for the knife” to “需要一塊磨刀石” (1384, expressing that the old man needs a stone for the knife). Furthermore, there are three modifications of the reporting clauses – changing the reporting clauses “he thought” to “他盤算好:” (1295, meaning he calculated well); “he thought that” in an IT (Indirect Thought) to the FDT “他想:” (1297, recreating a colon after the reporting clause); “the old man promised” to “老人暗下決心” (1320, meaning the old man was determined secretly). Also, Li seems to have a preference for rhetorical questions. There are two noticeable instances of changing the original statements from declarative to interrogative, one of which is changing “If I could have used a bat with two hands I could have killed the first one surely. Even now,” to “要是我雙手使棍,第一條肯定性命難保。誰笑我老了不中用了?” (1307-1308, meaning if I could have used a stick with two hands, the first shark definitely could have been killed. Who tease me for being old and feeble?) The rhetorical question “誰笑我老了不中用了?” emphasising the old man’s existing vigour is a creation that is not found in the original. Another instance of a creation of a rhetorical question is found in clause 1292 in which “I must not deceive myself too much” is rendered as “何苦自欺欺人?”, highlighting the old man’s affirmation not to deceive himself and others. In addition, there are five modifications made specially by Li to add concrete details of description regarding the part of the fish the shark tears (1293), the old man’s recall of his past accomplishment (1306), the healing effect of the salt water (1322), the way the old man puts the fish in the skiff (1323-1324) and the lights the old man wishes to see (1386-1388). Lastly, there are two restructurings of the sequences of details which appear to be creations of a marked syntactic structure in Chinese that are not attributed to the original; for example, the originally unmarked structure as it could be expressed as “(1)我現在(2)想些什麽好呢?”in the translation of clause 1304 is rendered specially into a marked structure by Li as “(2)我想些什麽好呢?(1)現在,” ; and “(1)人總不會在(2)海上迷路的, (3)何況島是長長的一條” as it could be rendered in the translation of clause 1332 is translated deliberately as “(2)海上,(1)人總不會迷路的,(3)何況島是長長的一條, ” emphasising respectively “the present moment” when the old man is thinking and “the sea” in which a man is never lost. Such creations of a marked syntactic structure produce special stylistic effects that are not found in the original.

Regarding Zhao’s translation, there are seven additions, six modifications and two restructurings of the sequences of details found. As for the seven additions made, they all appear to be minor details as logical presuppositions made explicit on the basis of the given information in the original to reproduce expanded descriptions; for example, adding the repetition “我決不要自己瞞自己” (1292, meaning I should not deceive myself) to “I must not deceive myself too much”. As for the six modifications, four are made to the reporting clauses of the FDT presentation. Overall, Zhao appears to be more creative than the other translators in rendering the reporting clauses “he thought” into varied descriptions of the old man’s “saying,” “estimating” and “asking” to himself instead of just merely “thinking” as that described in the original. Another creative rendering of Zhao’s translation is found in his translation of the deontic modalized expressions contained in the DS presentation– “He’ll be up soon and I can last. You have to last. Don’t even speak of it” (1302), which are rendered as “它快上來了,我撐得住。哼,你就得撐著,這還用說!”The interjection “哼” (pronounced as “heng”) is added to the original, giving the old man’s speech a livelier tone than that in the ST, and an exclamation mark “!” is added to produce emphasis at the end of the creation “這還用說!” , highlighting the old man’s self-esteem and confidence in his ability to endure the battle with the fish. Moreover, there is a slight adjustment made in the description of the old man’s sailing the skiff – “and he sailed the skiff to make his home port as well and as intelligently as he could” (1311), in which the circumstance of manner “as well and as intelligently as he could to” is changed to “駕得盡量穩當, 盡量用心,” depicting that the old man sails the skiff steadily and attentively, showing once again a more flexible rendering of the original diction “well” and “intelligently.”

Finally, there are two instances of the restructuring of the sequences of details found in Zhao’s translation of the deontic modals; one of which is in the modalized expression “Now I must prepare the nooses and the rope” (1288), in which the original order of “the nooses” and “the rope” is reversed as “繩子跟活套,” showing a slight adjustment of the original sequence of details. Another striking example of shift is found in the restructuring of the unmarked syntactic structure as it could be normally expressed in Chinese – (1)因為槳把子上有個把手, (2)要一隻手拿著才好使” – into a marked structure in the translation – “ (2)要一隻手拿著才好使,(1)因為槳把子上有個把手” (1305). In Chinese, the conjunction “因為” regarding the cause is normally put before “所以” regarding the result in the conjunctions linking clause, yet in this example Zhao deliberately follows the original sequence of events and presents the result “要一隻手拿著才好使” (He could only use it effectively with one hand) before the cause “因為槳把子上有個把手(because of the grip of the handle), reproducing a marked syntactic structure that appears to be influenced by the original. Such a way of recreating a marked syntactic structure is displayed similarly in Li’s rendering of clauses 1304 and 1332 as explained above. On the whole, Li and Zhao appear to be more flexible and creative in reordering the original sequences of events compared with Hai and Wu.

5.2 Translation of the Epistemic Modal Operators

Generally, most of the epistemic modal operators are retained in the four translations with only minor adjustments made by Hai and Zhao, and obvious shifts only to be found in Li’s version. Wu’s translation is still the most faithful to the original with no obvious shifts uncovered. There are totally four, seven, and eleven instances of the four types of adjustment found respectively in the versions of Zhao, Hai, and Liin the translation of the epistemic modal operators, which are all relatively less than those found in the translation of the deontic and boulomaic modal operators. Table 1.3 below first provides a list of the translations of the major epistemic modal operators identified in the four Chinese translations to show how they are generally rendered. It is then followed by Table 1.4, which outlines the frequencies of the four types of adjustment made in the translation of the epistemic modal operators in the four versions. As in the previous section, following the presentation of statistics is an overall description of the findings uncovered.


說不定, 也許可能, 沒準兒, 怕也未必, 或許


就會, 也許會, 該是, 還可, 會, 也許, 還會, 倒可以


也許, 作興, 或許, 沒準兒, 說不定


就會, 就能, 就, 會, 還要, 要, 也會

I wonder

我不懂, 不知道, 不明白, 我不曉得

I suppose

我猜想, 我看, 我想

should have

應該, 原該, 總該, 也會

might have

也許, 也許, 說不定, 那也可能

I do not know

我摸不透, 我弄不懂, 我不知道, 真不知道

I cannot know

我可沒法知道, 我沒法知道, 我沒法兒知道

Table 1.3     Chinese Translations of Major Epistemic Modal Operators in the The Old Man and the Sea Mini Corpus

Translation strategy / frequency




















Restructuring of the sequences of details





Total frequency





Table 1.4     Translation strategies used for  Epistemic Modal Operators

As shown in Table 1.4 above, most of the epistemic modal operators are rendered into their similar counterparts according to the original over the four translations. In Hai’s translation of the epistemic modal operators, there are slight adjustments made in the form of addition (a total of two instances) and restructuring of the sequences of details (a total of five instances), which are all contributed by the fronting of the reporting clauses “he thought” or “the old man thought.” As for addition, a reporting clause “他想:” is specially added to the FDT presentation – “I do not know. But I will try it once more” (1343-1344), while in the original the reporting clause “the old man thought” occurs only once at the beginning of the FDT presentation. Moreover, a slight addition is made by adding the expression “話又說回來” (meaning in retrospect) to the past subjunctive “But if I had, and could have lashed it to an oar butt, what a weapon” (1368-1369) to highlight the old man’s projected view on his possibility of lashing the shark to an oar butt. As for the five restructurings of the sequences of details, four of which are made by fronting the reporting clause “he thought” and translating it as “他想:” at the beginning of the FDT presentation, while one is contributed by fronting the reporting clause “the old man thought” and translating it as “老頭兒想:”. By fronting the reporting clause and adding a colon after it in every FDT presentation, Hai puts extra emphasis on the voice of the narrator in reporting the old man’s thoughts. Such a way of rendering the reporting clauses of the FDT presentation is found consistently throughout the corpus studied in Hai’s version as explained before.

As for Li’s version, one omission, one addition and nine modifications were found, accounting for a total of eleven shifts, showing that he has consistently made the most changes among the four translations investigated. Three of the nine modifications concern the epistemic modal auxiliaries “may” (1337, 1362) and “should have” (1359). The epistemic modal “may” in “It may make him jump through” (1337) is changed to “它該是要跳的呀” (meaning “it should jump”), shifting the low value of possibility conveyed by “may” in the original to the high value of certainty expressed by the modal auxiliary “該是”; and another “may” in “you may have much luck yet” (1362) is changed to “你交上好運的” (meaning “you will have luck”), thus similarly enhancing the value of possibility conveyed by “may” in the original to the value of probability conveyed by “會” in the translation. Moreover, the epistemic modal “should have” in “I should have brought a stone” (1359) is changed to “我帶一塊來就好” (meaning it would be good if I brought one), weakening the deontic sense of duty conveyed by the original deontic modal “should have.” In addition, the modal lexical verb “I wonder” contained in the FDT clause “I wonder how the great DiMaggio would have liked the way I hit him in the brain” (1353) is changed to the subjunctive “如果迪馬齊奧見我一叉擊中它腦門,不知該多麽高興?” in the form of an interrogative to denote the hypothetical situation that DiMaggio would be happy to see the old man hit the fish’s brain. The rhetorical question “不知該多麽高興?” (meaning “how he would have liked it”) is a creation that is not found in the original. Other than the modification of epistemic modals, there are two marked alterations of reporting clauses; one of which is changing the reporting clause “he had thought” contained in the IT clause “he had thought perhaps it was a dream” (1347) to “心裏就想過:”, shifting the original IT presentation to one of FDT by signaling the reporting clause with a colon, expressing the old man’s thought – “也許這只是個夢”(meaning “perhaps it is a dream”) in the first person from his own view. Another instance is found in the rendering of the reporting clause in the FDT presentation “and what he could do to a shark if he were swimming free. I should have chopped the bill off to fight them with, he thought” (1366-1367), in which the reporting clause “he thought” is fronted at the beginning of the FDT presentation and rendered specially as “他總想著大魚:” (meaning “he kept thinking of the fish”) with a colon after which to signal the presence of the narrator. Moreover, the subsequent detail regarding the old man who chops the shark’s bill in fighting with it is slightly modified as “我可以把它的劍吻砍下來當作武器” (meaning “I should have chopped the bill off and used it as weapon”). The remaining three modifications are all minor adjustments made to the concrete details such as that about the old man’s feeding many people (1358), and his confidence that the boy will worry about him (1364),

Regarding Zhao’s translation of the epistemic modal operators, only slight adjustments are made, including one omission, two additions and one restructuring of the sequence of details. For omission, the epistemic modal lexical verb “I am sure” in “But I am sure he would have confidence” (1364) is omitted, presenting the proposition more like a categorical assertion than a modalized expression as used in the original – “不過他一定會有信心。上點兒歲數的漁民,有好些會著急。” For the two additions, they are both further details added on the basis of the given information to explicate the contexts of events, such as adding “我撐不撐得下去” to “I do not know” (1342) to make clear that the old man does not know whether he can last or not, and adding the adverb “本來” (1378, meaning “originally”) to emphasise that the old man might originally have bought luck. Lastly, there is a restructuring of the sequences of events regarding the old man’s doubt about whether the great DiMaggio would have liked the way he hit the shark’s brain; the detail “would have liked the way,” translated as “喜不喜歡?” (1353), is placed specially at the end of the clause for emphasis, showing a more flexible rendering of the original sequences of events.

5.3 Translation of Speech and Thought Presentation

Regarding the translation of the speech and thought presentation, it appears that the strands of NRA, DS and FDT are generally rendered faithfully in the four translations, except for the special fronting of the reporting clause “he thought” or “the old man thought” and the addition of a colon after it to form the reporting clause “他想:” or “老頭兒想:” in Hai’s translation (with a total of fifteen instances, two of which are formed by a comma rather than a colon after each), and the recreation of the reporting clauses into a range of interesting expressions in Li’s translation such as “他盤算好:” , “他總想著大魚:” and in Zhao’s translation such as “他心裏在說,” “他心裏說,” “他估計” and “他心裏在問.”

FDT presentation is subject to important changes in the translations by Hai, Li and Zhao. The use of the colon after the reporting clause in Hai’s version and a few similar instances in Li’s translations as mentioned above can be said to highlight more strongly than the original the presence of the narrator in FDT. Furthermore, there are two peculiar instances in which the IT mode in Li’s translation is changed to FDT by recreating the reporting clauses “and he thought that” to “他想:” (1297), and “he had thought” to “心裏就想過:” (1347), presenting the old man’s thoughts from his own first-person view rather than from a third-person view via an external narrator in the form of IT as that appears in the original.

6. Conclusion

In view of the linguistic findings of the translation shifts gathered in this corpus-based study with regard to modality, it is clear that Li’s and Zhao’s translations have made more drastic changes to the original than Hai’s and Wu’s translations. Li has consistently made the most adjustments whereas Wu the least in the translation of all types of linguistic features investigated. Li uses rhetorical questions, changes declaratives into interrogatives, and modifies the reporting clauses creatively. In addition, the most prominent strategies Li uses are omission and modification. Li tends to omit rather than to explicate information, reproducing relatively reduced pictures of descriptions. Furthermore, he employs modification extensively to transpose the original images into new ones. Different from Li’s version, Zhao’s is the most creative among the four in reordering the sequences of information flexibly to reproduce different emphases or stylistic effects in descriptions. Restructuring of the sequences of details and addition are his outstanding strategies used. The additions are mostly presuppositions or inferences made explicit to explicate the specific contexts in which events take place. Furthermore, modifications are occasionally used to reproduce slightly shifted images. Yet compared to Li, Zhao makes less drastic adjustments; his translation replicates the original to a greater degree. It is worthy of attention that Li’s and Zhao’s translations both display new linguistic and stylistic features that are not found in the original and the other two versions, particularly in the rendering of FDT and the choice of diction. For example, they both invent marked syntactic structures, producing special stylistic effects that attract attention. Also, they both render the reporting clauses interestingly into varied expressions, representing the old man’s monologues in a livelier manner. In contrast to Li’s and Zhao’s translations, Wu’s translation is the closet to the original with only very slight adjustments made. As for Hai’s translation, except for generally fronting the reporting clauses “he thought” and translating them as “他想: ” to indicate the presence of the narrator, an instance of changing the FDT to DT, and a few instances of modifying the reporting clauses to highlight the old man’s strong determination to fight the battle, Hai’s translation can be considered to stick quite closely to the source text. More importantly, though obvious shifts occur in all four translations and they each exhibit their own distinctive features, generally they still fall within the boundary of the original in replicating it quite closely. Overall, the adjustments made in the translations by Hai, Wu and Zhao can be considered light. In addition, the NRA, the DS and the FDT presentation are generally retained with innovative creations of new stylistic features by Li and Zhao.

As a conclusion, I identified the recurring patterns of linguistic behavior or stylistic traits of the four translations of The Old Man and the Sea through systematic comparison of the original and its four translations by using simple corpus processing tools. The quantified findings generated serve as a useful starting point for spotting interesting areas of investigation, but they tend to be individual decontextualized lexical choices that provide little insight into the translators’ motivations. Thus, in future research, qualitative analysis could be conducted on the basis of the prominent findings of the quantitative analysis to inspect individual options within their immediate cotext and context. Those instances revealing distinctive individual styles of the four translators in the translation of modal expressions and speech and thought presentation could be selected for further close critical analysis. Moreover, extra-textual information related to the translators’ intention and orientation and the translation situations could also be collected for complementing the quantitative method of analysis in understanding better the translators’ choices.


Hai, G. (Trans.). (1957). The Old Man and the Sea [Lao Ren Yu Hai]. Shanghai: Xin Wenyi.

Han, Y. (2005). Introduction to Semantics and Pragmatics (Distance learning course materials). Hong Kong: The Open University of Hong Kong.

Hemingway, E. (1952, 1993). The Old Man and the Sea. London: Arrow Books.

Leech, G. N. & Short, M. H. (1981) Style in fiction. London: Longman.

Li, X. Y. (Trans.). (1987a) The Old Man and the Sea [Lao Ren Yu Hai]. Sichuan: Sichuan Wenyi.

Simpson, P. (1990). Modality in literary-critical discourse. In W. Nash (Ed.), The writing scholar studies in academic discourse vol. 3 (pp.63-94). London et al.: Sage Publications.

Simpson, P. (1993). Language, ideology and point of View. London & New York: Routledge.

Wu, L. (Trans.). (1987). Torrents of Spring / The Old Man and the Sea [Chun Chao / Lao Ren Yu Hai]. Shanghai: Shanghai Yiwen.

Zhao, S. W. (Tran.). (Tran.). (1987). The Old Man and the Sea [Lao Ren Yu Hai]. Lijiang: Lijiang.


[1]Examples are selected from the parallel corpus of The Old Man and the Sea I compiled to illustrate the specific linguistic items introduced. The number in brackets beside each example given refers to the number of the clause labelled originally in the corpus.

About the author(s)

Elaine Ng received a PhD in Comparative Literature from University College London in the UK. Her areas of research are stylistics, systemic functional grammar of English and Chinese, corpus-linguistics and translation studies. She is currently a lecturer in English Language at the Center for Language Education of Southern University of Science and Technology in Shenzhen, China. Previously, she served as an Assistant Professor and the Programme Coordinator of Applied Translation Studies at United International College (UIC) in Zhuhai, China from 2010 to 2013 where she taught English language and literature and translation courses. Earlier in her career after graduation from her PhD studies in 2009, she was a lecturer at the School of Continuing Education (SCE) of Hong Kong Baptist University. She has substantial experience teaching English, linguistics, translation and literature courses in Universities in Hong Kong and China. She has been actively participating in Conference and scholarly activities and has published peer-reviewed articles related to her research interests.

Email: [please login or register to view author's email address]

©inTRAlinea & Elaine Y.L. Ng (2017).
"A corpus-bases study of The Translation of Point of View in four Chinese Translations of Hemingway’s The Old Man and the Sea"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2255

Intra-language differences and translation quality assessment

An exploratory study on a learner corpus of literary translations

By Rudy Loock (Université Lille-Nord de France & UMR "Savoirs, Textes, Langage" du CNRS, France)

Abstract & Keywords

The aim of this article is to consider the question of whether the measurement of intra-language differences between original language and translated language can be used as a tool for translation quality assessment. To ask such a question is to enter the complex debate on the interpretation of intra-language differences: should we consider translated language as variation comparable to dialectal or geographical variation or should we consider that the over-representation or under-representation of a given linguistic feature means that the quality of the translation should be improved? Through the analysis of a learner corpus (translation tasks from English to French performed by advanced students enrolled in a translation master’s programme) for two linguistic features known to be problematic for English-French translation trainees (derived adverbs and existential constructions), our exploratory study aims to examine whether or not some correlation can be found between the observed intra-language differences and the overall quality of the translation tasks. The genre to be considered here is contemporary fiction texts.

Keywords: translation quality assessment, corpus-based translation studies, third code

©inTRAlinea & Rudy Loock (2017).
"Intra-language differences and translation quality assessment An exploratory study on a learner corpus of literary translations"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2258


The aim of this article is to consider the question of whether the measurement of intra-language differences between original language and translated language can be used as a tool for translation quality assessment. To ask such a question is to enter the complex debate on the interpretation of intra-language differences: should we consider translated language as variation comparable to dialectal or geographical variation or should we consider that the over-representation or under-representation of a given linguistic feature means that the quality of the translation should be improved? From an even more general perspective, should we consider that translated language is intrinsically different and represents a “third code” (Frawley 1984) or should we consider that linguistic homogenization (or absence of intra-language differences) between original and translated language goes hand in hand with translation quality, alongside the idea that in translation, “the utopian goal is to make it virtually impossible to tell the translation from an original text in that language” (Teubert 1996: 241)? Although the existence of the third code may be accounted for by translation universals and/or source language interference (see below), no consensus exists on the possible link with translation quality. Through the analysis of a learner corpus (translation tasks from English to French performed by advanced students enrolled in a translation master’s programme at the University of Lille) for two linguistic features which are known to be problematic for translation trainees (derived adverbs and existential constructions), our exploratory study aims to examine whether or not some correlation can be found between the observed intra-language differences and the overall quality of the translation tasks. The genre to be considered here is contemporary fiction texts.

The article is organized as follows. In the first part, we discuss the starting point of our study, namely the existence of intra-language differences and the two types of explanation that have been put forward in the literature to account for their existence: translation universals, which in their strictest sense exclude interference (see for example Baker and Olohan 2000, Olohan 2003), and source language interference (see for instance Rabadán, Labrador and Ramon 2009, Cappelle and Loock 2013). The second and third parts elaborate on the aim and methodology of the study. The fourth part is dedicated to the results of our two studies, (i) for inter-language (or cross-linguistic) comparisons between original English and original French in general, (ii) for comparisons between the English original texts and their translations in French by translation students in our learner corpus. The final discussion then concerns intra-language differences between original and translated French, taking into account the linguistic characteristics of the English original texts, with the final aim of checking whether or not a correlation exists between these results and overall translation quality. We also provide suggestions for further research, as this is meant to be a pilot study as part of a larger project on corpus-based translation evaluation.

1. The starting point: the third code

With the advent of corpus-based translation studies, which could be summed up as the use of the tools of corpus linguistics in descriptive translation studies (Laviosa 2002), many studies have provided evidence in support of what Frawley (1984), who sees translation as a recodification, calls the “third code” and defines as “a code in its own right, setting its own standards and structural presuppositions and entailments, though they are necessarily derivative of the matrix information and target parameters” (Frawley 1984: 169), supposedly different from the source and the target language. Such studies have shown that intra-language differences are a reality: translated language is linguistically different from original language. Following Baker’s (1993) seminal paper, translated language has been considered to be an object worthy of research in itself, and many researchers have uncovered the existence of intra-language differences with original language, sometimes independently of the source language(s) of the translations.[1] For instance, studies on translated English include Baker and Olohan (2000) and Olohan (2003), which have respectively shown, through the comparison between samples of original English and of English translated from a variety of languages, that the use of that is more frequent after reporting verbs (say, tell, report…) in translated English and that contractions are more frequent in original English, in a range of genres. Using a similar methodology, Laviosa (1996, 1997, 1998, 2002) has shown that translated English texts have a relatively lower percentage of content words as opposed to grammatical words (lower lexical density) as well as a higher rate of repetition for the most frequent words (lower lexical variety), as opposed to what is to be found in original English. Xiao (2010) has found similar results to Laviosa’s for Chinese. Jiménez Crespo (2010) has found that the syntactic subject is more often expressed in translated Spanish than in original Spanish – a pro-drop language where the subject does not need to be overt. Working on translations from only one source language, other researchers have also confirmed the existence of intra-language differences: for instance, working on English and French, Cappelle and Loock (2013) have shown that existential constructions (that is there-constructions in English and il y a-constructions in French, see below) are more frequent in French translated from English than in original French, but also less frequent in English translated from French than in original English. Rabadán, Labrador and Ramon (2009) have shown that adjectives are more often preposed in Spanish translated from English than in original Spanish.

What is shown by all these studies – which only represent a sample of studies on intra-language differences between translated and original texts that have been conducted over the last twenty years (see Laviosa 2002, Olohan 2004, Zanettin 2012, or Loock 2016 for further examples) – is that translated texts, whether from one or a range of different source languages, are linguistically different from original texts written in the same language. There is therefore a general consensus on the existence of intra-language differences. In fact, it seems that these differences are so systematic that translated texts can now be detected automatically by computers (see Baroni and Bernardini 2006 and Kurukowa, Goutte and Isabelle 2009). What researchers seem to disagree on, however, is how to account for such differences, as two types of interpretation can be found in the literature. On the one hand, some researchers claim that the differences can be explained with Translation Universals (TUs) as defined originally by Baker: “features which typically occur in translated text rather than original utterances and which are not the result of  interference from specific linguistic systems” (Baker 1993 : 243). Baker suggests the following are potential translation universals:

  1. explicitation (“overall tendency to spell things out rather than leave them implicit”),
  2. simplification (“tendency to simplify the language used in translation”),
  3. normalization/conservatism (“tendency to exaggerate features of the target language and to conform to its typical patterns”), and
  4. levelling (“tendency of translated text to gravitate towards the centre of a continuum”).

Other TUs have been added, such as Tirkkonen Condit (2002, 2004)’s unique items hypothesis, according to which a linguistic item that is specific to the target language is under-represented in translated texts. Other suggested candidates are disambiguation and avoidance of repetition (see House 2008 for a critical list). Examples of studies accounting for intra-language differences with translation universals are Baker and Olohan (2000) or Jiménez Crespo (2010), where, the over-representation of that after reporting verbs in English or of the syntactic subject in translated Spanish are seen as the results of explicitation.

On the other hand, some researchers, generally working on translations from one source language, claim that intra-language differences are due to source language interference (SLI), that is, they claim that the linguistic characteristics of the source language have an influence on the linguistic characteristics of translated texts and must explain the differences that can be observed with original language. For instance, because existential constructions are more frequent in original English than in original French, Cappelle and Loock (2013) claim that the intra-language differences that they observe between translated and original texts for English and French “strongly suggest source-language interference” (Cappelle and Loock 2013: 268). It is important to note however that source language interference has been listed as a potential TU (cf. Toury’s 1995 Law of Interference), but TUs à la Baker clearly exclude it (cf. definition above) and most studies on TUs do not mention interference as one of them (see Mauranen 2004 for a discussion of the unclear status of interference in relation to translation universals). If one considers interference as one of the universal “tendencies”, then the issue raised here can be reformulated as determining the influence of source language interference as opposed to other universals like simplification or normalization.

Some of these researchers (for instance Rabadán, Labrador and Ramon 2009, Loock, Mariaule and Oster 2014) then claim that a correlation may exist between intra-language differences and translation quality, equating linguistic homogenization and quality. For instance, Rabadán, Labrador and Ramon 2009: 323) claim that:

The smaller the disparity between native and translated usage in the use of particular grammatical structures associated with specific meanings, the higher the translation rates for quality.

2. Aim of the article

In this article, our goal is to try and see, through the analysis of a learner corpus containing students’ literary translation tasks, whether such a relationship as posited by Rabadán, Labrador and Ramon (2009) between linguistic homogenization and translation quality can be confirmed. Focusing on two linguistic features, derived adverbs and existential constructions, and on English to French translations by postgraduate students who are native speakers of French, we aim to measure the under-/over-representation of these two linguistic features in the translated texts, which have been the object of a prior, independent evaluation by a professional translation trainer, and see whether there is a correlation between this evaluation and the measurement of intra-language differences.

The two linguistic features that have been chosen for this exploratory study are (i) derived adverbs and (ii) existential constructions, which are both known to be problematic for translation trainees translating from English to French. Each of these will be considered in turn as a tertium comparationis for our comparisons, first between original English and original French in general (inter-language comparison) through the use of comparable corpora, and second between English original texts and their translation in French (also an inter-language comparison, but with the aim of discussing intra-language comparisons) through the use of our parallel learner corpus.[2]

The first linguistic feature, derived adverbs, corresponds to adverbs obtained through the addition of a bound derivational suffix to an adjective, –ly in English (finally, properly, honestly, quickly, slowly) and –ment in French (finalement, proprement, honnêtement, rapidement, lentement). The two sets of derived adverbs are generally presented as a case of translational equivalence (see Bertrand 1986), with the exception of morphological and semantic constraints, that is, cases of adverbs which have no equivalents in the other language (for instance successfully, familialement vs. *successeusement and *familially, which are lexical gaps) and false cognates, partial or full (for instance actually/actuellement; eventually/éventuellement). The second linguistic feature is existential constructions, that is, there-constructions in English and il y a-constructions in French, both structures having the same discourse function, that of introducing new elements into the discourse (see Bergen and Plauché 2005: 23 and Lambrecht 1994: 178 for a cross-linguistic comparison):

(1)       a. There is a dog in the garden.

            b. Il y a un chien dans le jardin.

This means that English –ly adverbs and French –ment adverbs as well as English and French existential constructions can be considered to “convey the same ideational and interpersonal and textual meanings(James 1980: 178) and to be translationally equivalent. However, in spite of their translational equivalence, the two linguistic features show a highly significant inter-language/cross-linguistic difference in frequencies between English and French (see section 4.1). In particular, derived –ment adverbs have traditionally been considered to lead to poor style (see for instance Ballard 1994: 100; Chuquet and Paillard 1987: 154-155; Vinay and Darbelnet 1977: 126, but this idea is to be found in many other translation textbooks). Because of significant differences in frequencies between original English and French, derived adverbs and existential constructions are generally seen as translationese-prone phenomena, that is, they represent linguistic features that could lead to some interference between original texts and translated texts with an over-/under-representation of the linguistic features (in this case, an over-representation in translated French and an under-representation in translated English).

3. Methodology

For this exploratory study we used two types of corpora. First, to measure inter-language differences between original English and original French, we used samples of literary texts extracted from the 100-million-word British National Corpus (BNC) via the interface set up by Mark Davies (http://corpus.byu.edu/) and from Frantext (http://www.frantext.fr), an electronic corpus containing ca. 250 million words of French literary, philosophical, technical and scientific texts from the 12th to the 21st centuries (see details of the samples below). To ensure that our two samples were comparable in the technical sense of the term we focused on post-1980 fiction. The details of these two studies are provided in Cappelle and Loock (2013) for existential constructions and in Loock, Mariaule and Oster (2014) for derived adverbs respectively. In this article, due to lack of space, we only provide the results without any methodological explanations.

Second, we compiled a learner corpus of translations performed by advanced students from English into French, French being their mother tongue. These students were all enrolled in the MéLexTra Translation Master’s programme in the English department at the University of Lille in France. In the first year of the programme, the students’ final assignment is the translation of a short story or a chapter from a novel from English into French. All texts belong to contemporary fiction. We selected 16 translation tasks, which had already been evaluated independently by the students’ translation trainer, who performed a 3-group classification: the best translations (group A), generally correct translations (group B), translations that can be improved (group C). The criteria used for this evaluation were fidelity to the original text as well as fluency of the target language. We collected both the original and translated texts for this learner corpus, whose size amounts to ca. 400,000 words. Table 1 provides a summary of the contents of the learner corpus; Table 2 provides the translation trainer’s 3-group classification.


Original English texts

Translated French texts

Minimum number of words



Maximum number of words



Average number of words



Total Number of words



Table 1. Content of the learner corpus


(good translation)


(correct translation)


(translation that can be improved)


(50,964 words)


(110,738 words)


(45,764 words)

Table 2. Independent 3-group classification by translation trainer

The methodology was as follows. First we checked the inter-language difference between English and French for the two linguistic features that we had selected using the comparable corpus which was set up using the BNC and Frantext. Then we analyzed our learner corpus of English>French translations and measured the frequencies of each feature in the original texts and in the translated texts, first globally (all translation tasks) and then for each student individually. We finally compared the extent of the difference in frequencies between original and translated texts with the 3-group classification established by the students’ supervisor to see whether there was a correlation between the two.

4. Results

4.1. Inter-language differences between English and French

In this section, we briefly provide inter-language results for original English and original French based on a comparison between the British National Corpus (BNC) for original English and Frantext for original French. Table 2 provides a summary of the samples that were analyzed.


Existential constructions

Derived adverbs

BNC (original English)



Frantext (original French)



Table 2. Contents of the comparable corpus for inter-language comparisons[3]


The two linguistic features show a highly significant difference in frequencies: in both cases, the feature is much more frequent in English than in French. Cappelle and Loock (2013) and Loock, Mariaule and Oster’s (2014) results are summarized in Figures 1 and 2.


Figure 1. Existential there is vs. il y a constructions in original English and French (per million words), obtained from the BNC and Frantext (p < 0.0001)


Figure 2. Derived –ly adverbs vs. –ment adverbs in original English and French (per million words), obtained from the BNC and Frantext (p < 0.0001)

4.2. Inter-language differences between English original texts and their translations in French by students

In this section, bearing in mind the inter-language differences in frequencies discussed in the previous section, we report on the analysis of our parallel learner corpus containing original English texts and their translations in French by advanced students, first for derived adverbs, then for existential constructions.

4.2.1. Results for derived adverbs

When we compare the frequency of –ly adverbs in the original texts with that of –ment adverbs in students’ translations, we can see that the overall inter-language difference discussed above is also present in our learner corpus. Figure 3 below shows the minimum, maximum and average frequencies per million words (pmw). What the results show is that students use fewer –ment adverbs in their translations than original authors use –ly adverbs. However, such general results are limited in that there is a lot of variation as is shown in Figure 3 and also, these results do not say anything about the systematicity of the pattern. This is why it is important to consider individual results, that is, to consider each translation task individually. These results are provided in Figure 4 and show that the pattern is systematic: all students’ translation tasks show a frequency for –ment adverbs that is lower than that of –ly adverbs in the original texts. Although we can see some variation concerning the extent of the difference between the frequencies in the original texts and those in the translated texts, it seems that the behaviour of students is homogeneous, reflecting the fact that in original French, derived –ment adverbs are less frequent than derived –ly adverbs in original English. Note however that the mean frequency for –ment adverbs in translated texts is much higher than that which is found in original French fiction texts (7662 vs. 4540 occurrences pmw) and that intra-language differences with original language therefore exist (the frequency in original literary French is indicated with a black dotted line in Figure 4). Only one translation task (number 9) shows a frequency of –ment adverb that is actually lower than the “norm” to be found in French original texts, but this is probably related to the low frequency of –ly adverbs in the original text in the first place.


Figure 3. General comparison in the learner corpus between the frequencies of –ly adverbs in English original texts (left) and –ment adverbs in the translated texts (right) (per million words)


Figure 4. Individual comparisons in the learner corpus between the frequencies of –ly adverbs in English original texts and –ment adverbs in the translated texts (per million words)

Because of the variation observed in the frequency of –ly adverbs in the original texts, which must have some influence on the linguistic characteristics of the translations, we then decided, before investigating the correlation with the A, B, C classification performed by the students’ translation trainer, to calculate differential results (in per cent) between each original text and its translation in French. For instance, student 1, with a frequency of –ly adverbs of 11,499 occurrences pmw in the original text and a frequency of –ment adverbs of 6538 occurrences pmw in the translation, has a differential result of -43 per cent. However, student 7 has a differential result of only 19 per cent, with a frequency of –ly adverbs of 14,635 occurrences pmw in the original text and a frequency of –ment adverbs of 11,823 occurrences pmw in the translation. It is these differential results that we tested for a possible correlation with the evaluation of the students’ supervisor. Figure 5 provides the differential result (in per cent) for each translation task.


Figure 5. Differential results (in per cent) between the frequency of –ly adverbs in the original texts and the frequency of –ment adverbs in the translated texts.

4.2.4. Results for existential constructions

With the same methodology as for derived adverbs, we investigated our learner corpus to find out the frequencies of existential constructions in original and translated texts. We first checked the general results, which are provided in Figure 6, and which show that the general pattern is in line with what one would expect based on the inter-language difference in frequencies between original English and original French (with existential constructions much more frequent in English). There is also a lot of variation in both original and translated texts. However, the pattern is not systematic when one considers results on an individual basis (see Figure 7). Note that on average, as for derived adverbs, the frequency of existential constructions in the students’ translation task is higher than that which can be found in French original texts (1079 occurrences pmw, shown by the black dotted line).


Figure 6. General comparison in the learner corpus between the frequencies of existential constructions in English original texts and in the translated texts (per million words)


Figure 7. Individual comparisons in the learner corpus between the frequencies of existential constructions in English original texts and in the translated texts (per million words)

This examination of individual results shows that the students’ behaviour is very heterogeneous: while some of them do use existential constructions with a lower frequency in their translations than the original authors do in the original texts, others actually use more existential constructions than what can be found in the original texts (students 1, 6, 9, 13, 16). The extensive increase between the number of existential constructions in the original text and in the translated text that can be seen for student 9 needs to be tempered, however, as the results for this student are skewed and need to be removed from our analysis: the over-representation of existential constructions in the translation can be explained by the repetition of Y’a pas le choix (lit. contracted form of ‘There’s no choice’) as a translation for You got no choice (10 occurrences in the original text). Results for student 9 are thus removed for the rest of the analysis.

As a consequence, the differential results are not all negative (note that student 9 has now been removed and the analysis rests on 15 translation tasks), as shown in Figure 8.


Figure 8. Differential results (in per cent) between the frequency of existential constructions in the original texts and in the translated texts.


4.3. Correlation with the independent evaluation

4.3.1. Derived adverbs

When we now divide these results based on the A, B, C categories and calculate the average differential result associated with each of the 3 categories, a very nice, regular pattern emerges for derived adverbs : group A is associated with the highest differential result, that is, the translation tasks in which the difference between the frequencies of original –ly adverbs and –ment adverbs in the translations is actually the biggest on average (-41 per cent), while group C is associated with the lowest differential results, that is, the translation tasks in which the difference between the frequencies of original –ly adverbs and –ment adverbs in the translation is actually the lowest on average (-32 per cent), with group B lying in the middle (-38 per cent). Figure 9 sums up those results.


Figure 9. Correlation between differential results for derived adverbs and independent 3-group classification

However, when considering the translation tasks in detail, we can see that the pattern is not so homogeneous (see Table 3). This means that in spite of the general correlation observed in Figure 9, one cannot consider that the correlation holds for each translation task. In other words, there is no direct relationship between an “overuse” of –ment adverb by students in their English-French translation task and the overall quality of this task; an overuse of –ment adverbs is not symptomatic of a poor performance (results range from -27 to -55 per cent for group A; from -19 to -47 per cent for group C, which is actually quite similar).

Group A

Group B

Group C




























Table 3. Details of the correlation between differential results for derived adverbs and independent 3-group classification

4.3.2. Existential constructions

With the same methodology, a correlation was sought between the differential results between the frequencies of existential constructions in original and translated texts and the A, B, C independent classification provided by the students’ translation trainer. As opposed to what was found for derived adverbs, no such general correlation exists: the highest differential results do not correspond to the best translations (group A) and the lowest differential results do not correspond to the translations in group C (see Figure 10). There is no need here to investigate individual results; the general results suffice to invalidate our hypothesis.


Figure 10. Correlation between differential results for derived existential constructions and independent 3-group classification

4.3.3. Discussion

The observation of the difference between the frequencies of the two selected linguistic features in the English original texts and their translation in French provides very mixed results and brings no concrete evidence of a correlation between linguistic homogenization and translation quality as far as the two linguistic features that we observed are concerned. This seems to suggest that the comparison of intra-language differences cannot be used as evidence for the quality of a translation task. We need to be very cautious here, as we are dealing with an exploratory study, but the analysis of our learner corpus suggests that there is no direct correlation between the frequency of use of a given linguistic feature and translation quality, even when the frequency of use clearly departs from the expected norm for original language.

It also seems that the behaviour of students is not similar and depends on the linguistic feature itself: a student who reduces the number of derived adverbs in the translation process will not necessarily reduce the number of existential constructions. Figure 11 clearly shows that students do not behave similarly for the two translationese-prone linguistic features that we retained for our study.


Figure 11. Differential results for both derived adverbs and existential constructions

Another limitation of our study is that we compare the frequency of a linguistic feature in each original text and in each translated text. But such an analysis actually provides no real information on the students’ translational behaviour: a reduced frequency of derived adverbs does not mean that students have only avoided translating the –ly adverbs in the original text with –ment adverbs in the target text; it is possible that they also added a number of derived –ment adverbs in some cases. The differential results cannot only be interpreted in terms of ‘omissions’ as they can also partly correspond to ‘additions’ (terms taken from Johansson 2007). This is something that we already noticed (Loock 2013) when we used both a comparable corpus and a parallel corpus to analyze existential constructions in professional translations. Overall results obtained by comparing frequencies in original and translated texts (such as those obtained in Cappelle & Loock 2013) actually hide more complex results; only the analysis of a parallel corpus can provide information on the omissions and additions of a specific linguistic feature by translators, as well as on translation strategies. The same approach could be adopted for students’ translations to analyze their translational behavior in more detail.

Finally, it is also important to note that the measurement of intra-language differences between original language and translations performed by professional translators also shows that frequencies for our two linguistic features also differ significantly with original language. In Cappelle and Loock (2013) and Loock, Mariaule and Oster (2014), it is clearly shown that such intra-language differences concern both French and English. We do not provide any data here but an analysis of post-1980 French translations from English fiction by professional translators (use of the CorTEx corpus of translated French from English fiction within the CorTEx project – Corpus, Translation, Exploration – and use of a self-collected parallel corpus of English-French contemporary fiction) shows that professional translators do use both derived –ment adverbs and existential constructions with a significantly higher frequency than that which can be found in original French. Similar results have also been found for French to English translations (based on the analysis of samples extracted from the Translational English Corpus), with lower frequencies this time in English translated from French than in original English for both linguistic features. We refer the reader to Cappelle and Loock (2013) and Loock, Mariaule and Oster (2014) for more information on the analysis of professional translators’ translations. If intra-language differences are to be correlated with translation quality as is claimed by the hypothesis which is formulated in Rabadán, Labrador and Ramon (2009) and which we tested in our exploratory study, then this means that translations performed by professional translators are also problematic, since they do not show linguistic homogenization either. Does this mean that there is room for improvement in these translations as well? Or is a certain range of deviation from the “norm” in original language acceptable? More generally, this raises the complex question of the third code: is it acceptable to say that translated language differs from original language (cf. translation universals as a natural phenomenon) or should translators aim to achieve linguistic homogenization between translated language and original language, meaning that third code is tantamount to translationese which as such should be eliminated from good quality translations? This complex question lies way beyond the scope of this article and has been the object of debates for a long time (cf. debates on the invisibility of the translator or the opposition between third code and translationese). Although our hypothesis in this article seems to suggest that we promote linguistic homogenization, we should remain aware that this is far from consensual.

5. Conclusions and future research

In this article, we have reported on an exploratory study on a learner corpus which investigated the possible correlation between, on the one hand, the frequencies of two translationese-prone linguistic features for translations between English and French and, on the other hand, translation quality. By comparing the frequencies of derived adverbs and existential constructions, first in a comparable corpus of original English and French, and then in a parallel learner corpus of English-French translation tasks, we have tested the hypothesis put forward by researchers like Rabadán, Labrador and Ramon (2009) according to which linguistic homogenization and translation quality go hand in hand.

In spite of providing interesting results, our exploratory analysis has not managed to completely confirm the hypothesis. Although results for adverbs do provide an interesting insight into the relation between an overuse of –ment adverb and translation quality, we have seen that such a correlation cannot be systematized for each translation task in our learner corpus. Results for existential constructions have proved to be completely inconsistent: a good translation is not directly linked to the frequency of the construction in the translated text in connection with that in the original text.

These results raise important methodological questions. First, our corpus is not very big and needs to be increased in size for future research. Second, we have used only one independent evaluator to categorize the translations into 3 groups: although this has the advantage of consistency, a second evaluator could confirm or contradict our evaluator’s classification. Also, the two linguistic features that we selected might not be the most discriminating ones: although they are known to be problematic for English-French students  who are native speakers of French (cf. translation textbooks like Chuquet and Paillard 1987, Ballard 1994), other linguistic features are known to lead to source language interference: -ing participle forms (vs. –ant participle forms in French), the use of the passive voice (more widespread in English than in French), the position of the adjective (overuse of prenominal position in French translated from English). By combining the results for a series of linguistic features, one might define a check-list of features to be checked in order to determine translation quality.

Finally, we have decided here to focus on the extent of differences in frequencies for two linguistic features in original texts and their translations. There is however another possibility: measuring the extent of the difference between the frequency of a given linguistic feature in a translated text and in original language as a whole. We decided against this option in this article because we wanted to take into consideration the linguistic characteristics of each original text – the texts might not be representative of original/translated language as a whole, as is shown by the variation that we observed. However, this might be another way of looking at the data, which might bring stronger evidence in favour of a correlation between intra-language differences and translation quality. This is currently being done to complement the results of this exploratory study, with translations performed both by students (our learner corpus described here, but with an increase in size) and by professional translators (use of samples from the TEC and of the CorTEx corpus).


Baker, Mona (1993) “Corpus Linguistics and Translation Studies: Implications and Applications” in Text and Technology, Mona Baker, Gill Francis and Elena Tognini-Bonelli (eds), Amsterdam, John Benjamins: 233-250.

Baker, Mona, and Maeve Olohan (2000) “Reporting that in Translated English: Evidence for Subconscious Processes of Explicitation?”, Across Languages and Cultures 1, no. 2: 141-58.

Ballard, Michel (1994) [1987] La Traduction de l’anglais au français, 2nd edition, Paris, Nathan.

Baroni, Marco, and Silvia Bernardini (2006) “A New Approach to the Study of Translationese: Machine-learning the Difference between Original and Translated Text”, Literary and Linguistic Computing  21, no. 3: 259-74.

Bergen, Benjamin K., and Madelaine C. Plauché (2005) “The Convergent Evolution of Radial Constructions: French and English Deictics and Existential”, Cognitive Linguistics 16, no. 1: 1-42.

Bertrand, Chantal (1986) “Quelques remarques sur les adverbes français en –ment et leur rapport sur les adverbes anglais en –ly”, Meta : journal des traducteurs 31, no. 2: 179-203.

Cappelle, Bert, and Rudy Loock (2013) “Is there Interference of Usage Constraints? A Frequency Study of Existential there is and its French Equivalent il y a in Translated vs. Non-Translated texts”, Target 25, no. 2: 252-75.

Chuquet, Hélène, and Michel Paillard (1987) Approche linguistique des problèmes de traduction anglais-français, Paris, Ophrys.

Frawley, William (1984) “Prolegomenon to a Theory of Translation” in Translation: Literary, Linguistic and Philosophical Perspectives, William Frawley (ed), Newark, University of Delaware Press: 250-63.

House, Juliane (2008) “Beyond Intervention: Universals in Translation”. Trans-kom 1: 6-19.

James, Carl (1980) Contrastive Analysis, London, Longman.

Jiménez-Crespo, Miguel A. (2010) “The Future of “Universal” Tendencies: a Review of Papers Using Localized Websites”, Talk given at the UCCTS 2010 conference, Edge Hill University, UK, 27-29 July 2010.

Johansson, Stig (2007) Seeing through Multilingual Corpora, Amsterdam/Philadelphia, John Benjamins.

Kurukowa, David, Goutte, Cyril, and Pierre Isabelle (2009) “Automatic Detection of Translated Text and its Impact on Machine Translation”, Proceedings of MT Summit XII,
URL: http://www.mt-archive.info/MTS-2009-Kurokawa.pdf (accessed 15 July 2014).

Lambrecht, Knud (1994) Information Structure and Sentence Form, Cambridge, Cambridge University Press.

Laviosa, Sara (2002) Corpus-based Translation Studies: Theory, Findings, Applications, Amsterdam, Rodopi.

Laviosa, Sara (1998) “Core patterns of lexical use in a comparable corpus of English narrative prose”, Meta 43, no. 4: 557-70.

Laviosa-Braithwaite, Sara (1997) “Investigating Simplification in an English Comparable Corpus of Newspaper Articles” in Transferre Necesse Est. Proceedings of the 2nd international conference on current trends in studies of translation and interpreting, 5-7 September 1996, Budapest, Hungary, Kinga Klaudy and Janos Kohn (eds), Budapest, Scholastica: 531-40.

Laviosa-Braithwaite, Sara (1996) The English Comparable Corpus (ECC): A Resource and a Methodology for the Empirical Study of Translation. Unpublished PhD diss, Department of Language Engineering, UMIST, Manchester, UK.

Loock, Rudy (2016) La Traductologie de corpus. Lille: Presses Universitaires du Septentrion.

Loock, Rudy (2013) “Close encounters of the third code: quantitative vs. qualitative analyses in corpus-based translation studies” in Interference and normalisation in genre-controlled multilingual corpora, Belgian Journal of Linguistics 27, Marie-Aude Lefer and Svetlana Vogeleer (eds): 61-86.

Loock, Rudy, Mariaule, Mickaël, and Corinne Oster (2014) “Traductologie de corpus et qualité: étude de cas”, Proceedings of the Tralogy II Conference, URL: http://lodel.irevues.inist.fr/tralogy/index.php?id=243 (accessed 15 July 2014).

Mauranen, Anna (2004) “Corpora, universals and interference” in Translation universals: Do they exist?, Translation Universals: Do they exist?, Anna Mauranen and Pekka Kujamäki (eds), Amsterdam/Philadelphia,  John Benjamins: 65-82.

Olohan Maeve (2004) Introducing Corpora in Translation Studies, London/New York, Routledge.

Olohan, Maeve (2003) “How Frequent are the Contractions?”, Target 15, no. 1: 59-89.

Rabadán Rosa, Labrador, Bélen, and Noelia Ramon (2009) “Corpus-based Contrastive Analysis and Translation Universals. A Tool for Translation Quality Assessment English-Spanish”, Babel  55, no. 4; 303-28.

Teubert, Wolfgang (1996) “Comparable or Parallel Corpora?”, International Journal of  Lexicography 9: 238-64.

Tirkkonen-Condit, Sonja (2004) “Unique Items – Over- or Under-represented in Translated Language?” in Translation Universals: Do they exist?, Anna Mauranen and Pekka Kujamäki (eds), Amsterdam/Philadelphia,  John Benjamins: 177-84.

Tirkkonen-Condit, Sonja (2002) “Translationese – a Myth or an Empirical Fact? A Study into the Linguistic Identifiability of Translated Language”, Target 14, no. 2: 207–20.

Vinay, Jean-Paul, and Jean Darbelnet (1977) [1958] Stylistique comparée du français et de l’anglais, Paris, Didier.

Xiao, Richard (2010) “How Different is Translated Chinese from Native Chinese? A Corpus-based Study of Translation Universals”, International Journal of Corpus Linguistics 15, no. 1: 5-35.

Zanettin, Federico (2012) Translation-driven Corpora, Manchester, St Jerome Publishing.


[1] In this article we use the term “intra-language differences” to describe linguistic differences between original and translated language. Although this may seem to be synonymous with Frawley’s “third code”, we aim to use “intra-language differences” as an objective term, while “third code” already is an interpretation for such differences (see below).

[2] We here refer to the distinction between parallel and comparable corpora, defined for instance by Teubert (1996) respectively as “a bilingual or multilingual corpus that contains one set of texts in two or more languages” and “corpora in two or more languages with the same or similar composition” (Teubert 1996: 245).

[3] The number of words in our samples differ, for the following reasons. First, the automatic retrieval of adverbs ending in –ment in the Frantext corpus requires the use of the tagged part of the corpus (‘Frantext catégorisé’), which is much smaller than the non-tagged part of the corpus, although for existential constructions the whole corpus was used. Second, a different methodological decision was taken for derived adverbs as opposed to existential constructions: Cappelle and Loock’s (2013) article on existential constructions include poetry and drama as fiction texts (ca. 250,000 words) while Loock, Mariaule and Oster’s study (2014) does not: the slight difference does not invalidate Cappelle and Loock’s (2013) study, but for derived adverbs we wanted our samples to be fully comparable and our learner corpus consists only of novels and short stories.

About the author(s)

Rudy Loock is Professor of English Linguistics and Translation Studies in the Applied Languages Department at the University of Lille, France, where he is in charge of the multilingual specialized translation master’s programme. His research interests include corpus-based translation studies, the use of corpora as translation tools, the didactics of translation, and translation quality. He is a member of the CNRS research lab “Savoirs, Textes, Langage”.

Email: [please login or register to view author's email address]

©inTRAlinea & Rudy Loock (2017).
"Intra-language differences and translation quality assessment An exploratory study on a learner corpus of literary translations"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2258

Issues in Computer-Assisted Literary Translation Studies

By Federico Zanettin (Università di Perugia, Italy)

Abstract & Keywords

In this article I first examine the application of computers to the study of literary texts and then provide an overview of how corpus linguistics tools and resources have been applied to the analysis of literary translation and translators. I discuss some of the issues involved in corpus design and construction, and consider some areas which may benefit from a corpus-based approach, including literary translation criticism and the study of translator style, one area that has received more attention than others. Finally, I suggest that quantitative and qualitative perspectives on computerized analysis of literary translation should converge and be integrated. 

Keywords: corpus-based translation studies, parallel corpora, stylometrics, corpus linguistics, literary translation

©inTRAlinea & Federico Zanettin (2017).
"Issues in Computer-Assisted Literary Translation Studies"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2254

1. Computer-assisted literary studies

Computer-assisted studies of literature started in the 1960s, taking advantage of the capability of computers to identify strings and patterns in electronic texts. With the help of computers, literary scholars could quickly access a list of all words in a text, and for each word access all the contexts in which that word appeared. While the use of electronic media as a means of scholarly analysis is still somewhat controversial among literary critics (Rommel 2004, Cohen 2010), digital humanities have been developing steadily over the years. Perhaps two main research areas can be discerned: the first revolves around the creation of electronic scholarly editions, thematic research collections and archives. Electronic versions of literary works are resources that facilitate access to texts by allowing for full text searches. Furthermore, the integration of metadata and multimedia in a dynamic hypertextual environment grants the researcher access to information about the text, for instance transcriptions and annotations from historical and scholarly editions, the reproduction of original sources such as manuscripts, and so on (Price 2008). A second area of investigation focuses on textual analysis within the theoretical framework of literary-linguistic stylistics: “(i)n this context, texts are seen as aesthetic constructs that achieve a certain effect (on the reader) by stylistic features on the surface structure of the literary text” (Rommel 2004: 89).

Approaches to computer-assisted stylistics vary: on the one hand, access to the text in electronic format is seen as a help to navigate and retrieve passages and citations in context in order, for instance, to explore imagery and characterization in a novel. The use of text analysis software is here functional to ‘traditional’, qualitative textual analysis, supplementing, corroborating and perhaps even challenging critical and interpretative reading of the whole text. On the other hand, more quantitatively-oriented approaches involve the use of statistical methods of analysis, based on frequency counts of any textual feature that can reliably be identified. Such studies are generally known as stylometry and have been successfully used in authorship attribution, both for legal purposes (forensic linguistics) and in literary-linguistic studies. They are based on the idea that each text bears the hidden signature of its author, the trace left by unconscious linguistic habits which is recoverable by comparing patterns in the use of textual features. As Hoover (2008: online) explains in his overview of quantitative approaches to literary studies,

Words …, as the smallest clearly meaningful units, are the most frequently counted items, and syntactic categories (noun, verb, infinitive, superlative) are also often of interest, as are word n-grams (sequences) and collocations (words that occur near each other). Thematic or semantic categories (angry words, words related to time), while more difficult to count, have the advantage of being clearly relevant to interpretation, and automated semantic analysis may reduce the effort involved. Phrases, clauses, syntactic patterns, and sentences have often been counted, as have sequences or subcategories of them (prepositional phrases, subordinate clauses, passive sentences). Many of the items listed above are also used as measures of the lengths of other items: word length in characters, sentence or clause length in letters or words, text length in words, sentences, paragraphs, and so forth.

One of the most successful procedures in authorship attribution was put forward by John Burrows (2002a) and is known as Burrows’s Delta. This simple but usually accurate measure of textual similarity and difference is based on the relative frequencies of the 150 most common words, and “is designed to pick the likeliest author of a questioned text from among a relatively large number of possible authors” (Hoover 2008: online).[1] Two further measures, introduced at a later date (Burrows 2006), are called Zeta and Iota, and focus on middle and low frequency words. Burrows applied his methodology to a number of research areas including translation, as will be seen in section 2.

Within translation studies, some research can be examined within the framework of literary translation criticism. As Paloposki (2012) explains, translation criticism can be understood as the (journalistic) practice of reviewing translations, as “related to prescriptivism or values and ideological choices” (Paloposki 2012: 185), or as a hermeneutic tool for the non-judgmental evaluation of literary translations. In this latter sense, “it explores the interpretative potential of a translation [looking] at degrees of similarity to or divergence from the source text’s perceived interpretative potential” (Hewson 2011: 6-7). Much research based on parallel corpora can indeed be seen in the light of such an approach. Some studies have focused on one parallel text, for instance Munday (1998), who uses a computerized version of a short story by Gabriel García Marquez and its German translation by Edith Grossman, and Mahlberg’s (2007), who looks at Gustav Meyrink's translation into German of Charles Dickens’ Bleak House. In most studies, however, one or more source text/s is/are accompanied by multiple translations by different translators, in what Johansson (2003) later termed “star” model parallel corpora.

One of the first studies to introduce a computer assisted study of literary translation was Maczewski (1996), who proposed the acronym CoALiTS (Computer Assisted Literary Translation Studies) to denote a field of research combining literary and linguistic computing with literary translation. He exemplified this approach through a case study, a computer-assisted analysis of the first chapter of Virginia Woolf's The waves, together with two French and three German translations. He created a small multilingual parallel corpus, which was linguistically annotated for lemma and part-of-speech, aligned at phrase level and analyzed through the use of the PALIMPSEST suite of programs written by the author himself. The application provides for parallel browsing and concordancing, and for deriving statistics about word and phrase order, which were used to compare the various translations in relation with the source text. Thus, Maczewski found that, while all German translations followed closely the wording of the English source texts, the two French translations differed more amply, one being “clearly target-oriented” and reading smoothly in French, the other giving priority “to the original's syntax and lexis” (Maczewski 1996: 183), and letting the source language “shine through” (ibid.: 184). Maczewski explained this partly as a consequence of closer structural similarity between English and German than between English and French, and partly as a result of different translation strategies.

Bosseaux (2001, 2004, 2007) investigated how translators manifest their ‘voice’, that is their discursive presence and point of view, in a number of French translations of two of Virginia Woolf’s novels, The Waves and To the Lighthouse. Using both monolingual and parallel concordances, Bosseaux considered features such as personal pronouns, time and space adverbials and verb tense in order to investigate deixis, modality, transitivity and free indirect discourse, and determine “whether and how the translator’s choices affect the transfer of narratological structures” (Bosseaux 2004: 108). Winters (2004, 2007, 2009) studied a number of style-related features in two different German translations (by Hans-Christian Oeser and Renate Orth-Guttman) of F. Scott Fitzgerald’s The Beautiful and Damned. Her analysis focused on modal particles, foreign words, code-switches and speech-act reporting verbs, and showed how the two translators differ significantly in their use of these features.

As noted by Huang and Chu (2014: 127), studies based on corpora containing different translations of the same source text have the advantage of detecting the influence of source text features on target realizations, but at the same time they define the concept of style in translation as a measure of difference from the source text. Thus, although they

demonstrate that individual translators can adopt quite different approaches to the translation of the same source text, their results do not reveal whether the patterns they identify are indeed consistent stylistic traits in each translator's work, rather than reflecting personal and circumstantial interpretations of a specific text. (Saldanha 2011b: 33)

Dorothy Kenny (2001) used a corpus of English translations of “experimental” novels by German authors to study lexical creativity in translation. She used the German-English Parallel Corpus of Literary Texts (GEPCOLT) in conjunction with two general reference corpora, one for German and one for English, to investigate whether translators resorted to lexical creativity to the same extent as source text authors, or rather showed a tendency towards normalization (one of the hypothesized universal features of translation, see below). Kenny looked at how hapax legomena (words only used once), writer-specific forms, and unusual collocations in the source texts are translated, and her results suggested an overall tendency towards normalization as concerns single words, and a combination of linguistic creativity and normalization– depending on individual translators – as concerns the translation of creative collocational clusters. Kenny was primarily interested in normalization as a general feature of translated texts rather than in the study of translator style, but her results also point to patterns of consistent variation in the linguistics habits of individual translators.

While corpus-based analyses of source texts and translations can be certainly carried out using plain text corpora, the enrichment of such corpora with lexical, syntactic and semantic annotation can in many cases facilitate or enhance the analysis. Furthermore, as stressed by Rommel (2004: 92) “[t]he importance of markup for literary studies of electronic texts cannot be overestimated, because the ambiguity of meaning in literature requires at least some interpretative process by the critic even prior to the analysis proper”. While much linguistic annotation can to a large extent be implemented automatically through existing NLP resources (Zanettin 2012), information regarding other textual features relevant to the analysis of literary translations can only be added through some degree of manual intervention. One interesting case, for instance, concerns the annotation of proper names. Van Dalen-Oskam (2013) tagged names in 10 English novels and their Dutch translations, and in 10 Dutch novels and their English translations arguing that, while onomastic research in literary stylistics has mostly been qualitative - focusing on the meaning of particularly significant names-, computer-aided quantitative onomastics may also provide interesting results. She found, for instance, that English translators from Dutch tend to add more name mentions (with respect to source texts) than Dutch translators when translating from English. Van Dalen-Oskam’s (2013) study highlights one important aspect of annotation: after trying to apply existing named entity recognition and classification (NERCs) tools to speed up the tagging of names, she discovered that they would not work properly with literary texts, and eventually had to resort to manual tagging. In order to be reliable, manual name tagging had to be carried out by Dutch and American scholars each working interpretatively with texts in their own language/culture, which required, in turn, an assessment of inter-coder agreement.

A type of annotation, which is specific to parallel corpora, is alignment annotation. In fact, while it is possible to compare features of source and target texts independently, for example, the frequency of specific words or the respective STTR, parallel corpora are best exploited when the researcher can access parallel segments using appropriate visualization tools. This is especially the case for qualitative oriented research which privileges the interpretation of individual instances, whose retrieval and analysis is facilitated by browsing and concordancing aligned parallel corpora. Various systems for the automatic alignment of parallel texts are now available, which reach a very high accuracy rate in mapping correspondences between “equivalent” segments and creating bitextual “translation units” (Zanettin 2012). However, literary texts seem to be especially “resistant” to alignment (Mikhailov 2002), making this a very time consuming enterprise, which also explains why parallel corpora of literary texts are a scarce commodity. The type of difficulties involved in breaking up a text pair to obtain parallel segments can be appreciated by considering a paragraph from a Salman Rushdie’s novel, Midnight’s Children, and its Italian translation (I figli della mezzanotte) by Ettore Capriolo (Zanettin 2001).

'A jewel,' he said, honking into a handkerchief, 'Sir and Madam, your daughter is a jewel. I am humbled, absolutely. Darn humbled. She has proved to me that a golden voice is preferable even to golden teeth.’

"Un gioiello," disse, strombettando nel suo fazzoletto. "Signore, signora, vostra figlia è un gioiello. Sono mortificato. Assolutamente mortificato. Mi ha dimostrato che una voce d'oro è preferibile persino ai denti d'oro."

Both an automatic aligning system and a human researcher would probably find it difficult to link together sentence/segment pairs from the above paragraphs. Based on punctuation, in fact, an automatic system may find four “sentences” in the English paragraph and five in the Italian one, corresponding to the number of sentence final full stops. The difference in segmentation between Italian and English can be mostly traced to different conventions in the rendering of direct speech: whereas in the source text the narrator’s comment is embedded in the first sentence of the dialogue, in the translation it separates that statement in two sentences. Additionally, the second and third sentences in the source text (corresponding to the third and fourth sentences in the translation) are very short, and they are closely interrelated through the repetition of the word “humbled”. Thus, they should perhaps best be interpreted as a single bitextual segment, which includes a full stop. This shows that the optimal size and boundaries of alignment units (i.e. bitextual segments) in parallel corpora of literary texts cannot be decided in advance on the basis of formal features, and that in order to fine tune the alignment for subsequent analysis of parallel concordances a great deal of manual editing may be involved. As a side benefit of this, however, manual alignment editing can also be seen as an essential analytical stage that helps the researcher to spot and categorize features worth investigating (in this case, for instance, patterns of repetition and parallel structures).

2. Corpus-based literary translation studies

Studies in corpus-based literary translation include translation criticism, but also general investigations of literary translation norms, such as Øverås (1998); the assessment of how differences in language and style between two translations of the same texts produced at different times relate to language change, as in Ji’s (2012) analysis of two late 20th century Chinese translations of Cervantes’ Don Quijote; the study of indirect translation envisioned by Zubillaga et al. (2015) when creating their trilingual parallel corpus of literary translations from German into Basque, with Spanish as a relay language. However, the majority of corpus-based studies on literary translation have perhaps been carried out in the context of research concerning “translator style”, defined as a coherent and motivated patterns of choices “recognizable across a range of translations by the same translator”, which “distinguish that translator’s work from that of other translators” and which “cannot be explained as directly reproducing the source text’s style or as the inevitable result of linguistic constraints” (Saldanha 2011a: 240). Different investigations have compared a range of features across various types of corpora. Quantitative procedures applied to the analysis of translator style refer to the frequencies and distribution of lexical or syntactic features, and are not dissimilar to those used in computational literary stylistics more generally, as outlined by Hoover (2008).

While most studies mentioned so far have centred their analysis on the comparison of parallel corpora of source texts and translations, Baker (2000) proposed that an approach based solely on a corpus of translations, such as the Translational English Corpus (TEC), could be used to investigate the style of literary translators. She suggested that by comparing between the sets of texts produced by different literary translators working in the same language from different source languages it may be possible to identify distinctive linguistic habits, manifested as consistent patterns of stylistic variation. Baker shifted the focus of the analyses away from the congruence between the style of the translations and that of the source text, onto individual literary translators, to investigate whether they can be shown to use distinctive styles of their own. She argued that corpus-based research should “explore the possibility that a literary translator might consistently show a preference for using specific lexical items, syntactic patterns, cohesive devices, or even style of punctuation, where other options may be equally available in the language” (Baker 2000: 248). Style is here defined, not as a way of responding to the source text, but rather, much as in literary-linguistic stylistics and forensic linguistics, as a “thumb-print” resulting from conscious or unconscious linguistic choices. In a tentative experiment, Baker (2000) applied to two sets of works by Peter Bush and Peter Clark (translating from Portuguese and Spanish, and from Arabic, respectively) some basic statistical measures that had already been used in other studies as indicators of translation universals. These are standardized type/token ratio (STTR) and average sentence length (used as indicators of simplification, Laviosa 1997, 1998) and frequency of reporting structures introduced by the English verb say (Olohan and Baker 2000). However, “although Baker finds interesting patterns in the work of both Peter Clark and Peter Bush, because her corpora do not include source texts it is difficult to prove conclusively that these patterns are not carried over from the source text” (Saldanha 2011b: 32). This point was later further investigated by Huang and Chu (2014), who applied Baker’s (2000) measures to a corpus of literary translations by two translators, both of whom translated from Chinese into English. They found little variation in the STTR and mean sentence length in the two sets of translations, concluding that these statistical measures do not seem to be reliable as indicators of translator style. Noting also that the distribution pattern of the reporting verb say in all its forms is similar for both translators and to the average for the TEC corpus, they suggest that the more marked differences in Baker’s study are probably due to the difference in source language, as Baker herself had suggested it may be the case. Zanettin (2000) shows how the source language variable could be kept under control by relating measures of STTR for different translators to the average STTR in large reference corpora of translations, for each of the source languages involved.

Huang and Chu (2014) also compared results for STTR, average sentence length and frequency of reporting verbs in a corpus of several translations by different authors from different source languages[2] with those from a corpus of several works by a few English writers, and found that there was much less variation among translations than among works originally written in English. A similar trend was also noted by Rybicki (2006), who analyzed two different English translations of three novels by the Polish author Henryk Sienkiewicz. By looking at the correlations between the relative frequencies of the most frequent words in dialogues, Rybicki was able to discriminate the characters’ idiolect, that is the words and patterns typically used by different characters. Multidimensional scaling plots for translations revealed to be very similar to those of the source texts, though at times the characters’ idiolects “are less different between translations than between either translation and the original” (Rybicki 2006: 102). While not useful towards the identification of traits of individual translator style, such results could be seen as pointing towards “leveling out” (Baker 1993) or “convergence” (Laviosa 1998) as a more general feature of (literary) translation.

Baker (2000) suggests that, while some of the differences between translators can be traced to the influence of source language features or to stylistic choices of the source text author, it may, nevertheless, be possible to identify target language patterns that are less likely to be affected by the source text. One such pattern may be the relative frequency of the optional that in English reporting structures, which can be seen as an indication of a smaller or larger degree of syntactic explicitation by different translators. In her corpus, she found a rather different distribution of zero/that realizations in the works of the two translators, apparently independent of the source languages. The same pattern is investigated by Saldanha (2008, 2011a, 2011b) who, however, in order to control the source language variable, compares works translated by two different translators (Margaret Jull Costa and Peter Bush) from the same source languages (Spanish and Portuguese), and combines Baker’s methodology with parallel corpus analysis. She finds that Jull Costa tends to use the optional that after reporting verbs with a frequency similar to that found in the reference corpus of translations (the TEC), while Bush’s usage exceeds the higher average frequency in a comparable corpus of English non translated literature (extracted from the British National Corpus). Saldanha also looks at the use of typographical markers such as italics and quotation marks, to investigate whether consistent patterns of choice differentiate the works by two translators, and if so whether these features can be shown to be due to the translators’ stylistic preferences or whether they are triggered by the source texts. She finds that, once the italics carried over from the source texts are subtracted,[3] one translator (Jull Costa) makes comparatively greater use of italics for emphasis, thus favouring readability in her translations, while the other (Peter Bush) makes a comparatively greater use of italics to highlight source language words. Saldanha also finds that, whereas Jull Costa uses italics mostly for emphasis and often accompanies non-italicized foreign words with glosses, Bush consistently uses italicized foreign words more often than the average in a reference corpus of English translations from Portuguese (extracted from the COMPARA parallel corpus, Frankenberg-Garcia and Santos 2003). Saldanha (2011b) concludes that these patterns of variation may be regarded as individual stylistic traits, which point to different explicitation strategies used by the two translators.

Zanettin (2001) used a parallel English-Italian corpus of six translations by two different translators (Ettore Capriolo and Vincenzo Mantovani) and their source texts by the same author (Salman Rushdie) to analyze the frequencies of Italian translations of expressions of indeterminacy such “a sort of”, “a kind of” and “something of a”, and found that most of the times the English expressions were translated by either one of two Italian phrases, namely "una sorta di" and "una specie di", with seemingly no preference for one or the other. However, and regardless of source text stimulus, it appeared that Capriolo has a strong preference for the phrase “una sorta di”, which he uses in about 70% of all cases, whereas Mantovani has a strong preference for “una specie di”, which he uses in almost 77% of all cases.

As opposed to studies based on parallel corpora of source and target texts, a number of experiments conducted within a target monolingual framework have applied Burrows’s Delta and other quantitative techniques to different analytical tasks. Burrows (2002b) himself tested his Delta methodology on a corpus of literary translations, by comparing a half a million word corpus of English Restoration poetry with the English versions of Juvenal's tenth satire by fifteen different translators, four of which were also authors of the works in the corpus. He found that, while Delta is not very successful when used to attribute the translations to specific authors, it manages to associate a given author/translator to the right translation in three cases out of four. That is, while it was not able to identify Dryden among all others as the author of one of the translations, knowing that Dryden is the author of one of them, it was able, nevertheless, to rightly identify the one most likely to be his. In a study that relies on quite sizable corpora of translations, Rybicki (2012) used Burrows’s Delta to compare between three sets of texts translated into and out of English, Polish and French. The corpus contained literary translations by different translators of multiple source texts by different authors. Cluster analysis by means of tree diagrams showed that works by the same author are grouped together, whereas no such pattern is visible for works by the same translator. Rybicki and Heydel (2013) experimented Burrows’s Delta with a case of collaborative translation involving a Polish translation of Virginia Woolf’s Night and Day. The translation was begun by Anna Kołyszko, who died in 2009 leaving the manuscript unfinished. Magda Heydel, a translator and Woolf scholar, took over editing the finished chapters and translating what was left. Through this method they were able to clearly distinguish which chapters were translated by Kołyszko and which by Heydel, who was consulted to confirm chapter attribution. Conversely, Hung et al. (2010) were able to show, using principal component analysis and other statistical techniques, that some medieval Chinese translations of Indian Buddhist texts traditionally attributed to different translators, were in fact translated by the same translator or group of translators.

Ji and Oakes (2012) used a range of statistical tests (Student t test, chi-squared test, Pearson’s r, etc.) to look at stylistic variation in a corpus containing two late 19th century English translations (by Bowra and Joly) of the Chinese classic novel Hongloumeng by Cao Xuequin. They considered both syntax and lexis, and compared features such as words and sentence length, part-of-speech distribution, words expressing “emotion” and “value”, fixed-phrases and n-grams, without taking into consideration the source text. Their results suggested that “when compared to Bowra’s earlier translation, Joly’s version … has deliberately intended to enhance the idiomaticity of the translated language through an idiosyncratic use of English idiomatic expressions and terms” (Ji and Oakes 2012: 205-206). Finally, Wang and Lee (2012) compared two different Chinese translations (by Xiao and Jin, respectively) of James Joyce’s Ulysses together and with the source text, and then one of the two translations was compared with two subcorpora by the same translator/author, one of translations and the other of original writing. First, they compared the wordlists from the two translations, finding that one translator (Xiao) tends to use colloquial verbs and particles more frequently than the other. Furthermore, they observed that Xiao’s idiosyncratic use of lexicon in translation mirrors the same phenomenon in his original writing.

While much corpus-based translation studies research has used literary narrative texts as primary data, quite often these corpora were not concerned with literary properties, analysis or criticism. Rather, they are used to explore the features of translation in a given language or of translated language as a variety in itself, as a result of the translation process and independently of variables such as source language, date and text type or genre. Thus, for example, while two of the first empirical studies based on corpora of translations (Gellerstam 1986, Laviosa 1997) contained exclusively or almost exclusively literary fiction, the purpose of the investigation was linguistic rather than literary. Gellerstam compared a corpus of translated Swedish fiction with a corpus of fiction originally written in Swedish in order to see whether computer-aided analysis could uncover “translation fingerprints” (Gellerstam 2005: 202). Laviosa (1997, 1998), instead, first used the TEC as a test bed for investigating some hypothesized universals of translation, i.e., supposedly invariant features, which characterize all translated texts, irrespective of source text and language (Baker, 1993). Results from these corpora as often presented as provisional evidence in (dis)favour of the “translation universals” hypothesis, needing to be corroborated by congruent data from larger, more varied, translational corpora in different languages.

3. Features of translator style

According to Hoover (2008: online), when investigating style,

[t]he need for some kind of comparative norm suggests that counting more than one text will often be required and the nature of the research will dictate the appropriate comparison text. In some cases, other texts by the same author will be selected, or contemporary authors, or a natural language corpus. In other cases, genres, periods, or parts of texts may be the appropriate focus. Counting may be limited to the dialogue or narration of a text, to one or more speakers or narrators, or to specific passages.

Studies of translator style have compared different translations of the same source text(s), translations and source texts by the same or different authors, translations by the same authors from the same or from different source languages, translations and non translations by the same writers, translations and reference corpora of non translations in the same language, and source texts and reference corpora in the source languages. In order to compare the work of individual translators with a norm representing a wider range of uses, one or more reference corpora have been used together with the main corpus, this latter containing either different sets of translations or translations and source texts. Corpora used for reference include corpora of non-translated texts in the source language/s, corpora of non-translated texts in the target language/s, and corpora of translated texts in the target language/s.

Huang and Chu (2014) propose a distinction between features that refer to the translator’s “conscious, purposeful and consistent” way of “transferring the ST features to the TT” (Huang and Chu 2014: 136), and features that focus “on the habitual linguistic behaviour of individual translators [and] results from linguistic patterns that are the translator’s subconscious choices.” (ibid.). According to them, these two types of translator style, which they label “S-type” and “T-type”, should be investigated using different methodologies, to be then combined into an integrated model. Thus, first “S-type” style should be detected starting from the comparison of multiple translations of the same source text, and then the consistency of each translator’s style should be verified by using comparable corpora composed of all their translations respectively.

However, not all combinations and comparisons are possible in all cases since, for example, not all texts are translated more than once, not all translators are also authors of original works in the same genre, nor do they all translate from more than one language and, more pragmatically, because not all types of corpus resources may be available for a given piece of research. Furthermore, while each of these types of comparison, usually in some kind of combination (e.g. one source text, multiple translations of it, a reference corpus of translations from the same language), adds a fragment to the general picture, on the one hand increasing the number of authors, languages and source texts makes it more difficult to control target language variables; on the other, increasing the number of translators, translations and non-translated texts in the target language makes it more difficult to control source language variables.

The linguistic patterns investigated rely both on quantitative and qualitative features, and, hence, motivated individual preferences are manifested both as quantitative, largely unconscious linguistic patterns, and as consistent and usually conscious translation choices. Measures of difference, based in some cases on simple word counts and in others on quite complex statistical methods, have taken into consideration the frequency of surface features such as orthographic words (as in STTR and Delta), or features derived from annotation regarding, for instance, parts of speech and words expressing “emotion” and “value” (Ji and Oakes 2012), or typographic features such as italics and quotation marks (Saldanha 2011a, 2011b). The analysis of these features is sometimes totally automated, and sometimes partially automated based on manual classification and selection of relevant instances, for example, by filtering out occurrences of reporting verbs which do not require optional that (Olohan and Baker 2000), or classifying different functions of italics (Saldanha 2001a, 2011b).

Summing up, the results of existing literature suggest that, while the style of a translated text is influenced to a large extent by that of the specific source text, author and language being translated, individual translators, nevertheless, do have their own distinctive style. In fact, some linguistic features have been shown to be independent of source text variables, and may be thus traced to individual preferences. It is possible to attribute “translatorship” of different parts of the same texts to two different translators, or to attribute one out of many translations of the same source text to a given translator, and even to identify patterns of linguistic behaviour which vary consistently across translators, while not being triggered by source text/language features.

While some measures of difference between texts produced by distinct translators seem to be, to a large extent, independent of the translation activity (i.e. they also characterize a translator’s non-translational writing activity), others can be, and have also been, interpreted as indicators of hypothesized universal features of translations. STTR, for instance, has been investigated as an indicator of lexical simplification (Laviosa 1998), identifying the optional reporting that in English as an indicator of explicitation (Olahan and Baker 2000), and the distribution of unusual collocations as an indicator of normalization (Kenny 2000). This can be interpreted as meaning that not all translators behave in the same way, in line with Toury’s view of “translation universals” as probabilistic tendencies rather than deterministic laws (Toury 2004). Thus, within an overall trend towards normalization, some translators may normalize more than others, staying below or above the average values for translated texts as regards, for instance, the relative frequency of unusual collocational patterns, compared to that in non-translated texts. Consistent patterns of choice regarding these features, which may be the conscious or unconscious product of specific translation strategies, can thus be seen as indicators of different individual stylistic traits. Incidentally, the results of two studies which set out to investigate translator style using target monolingual corpora seem to confirm a tendency towards one less studied hypothesized universal feature, that is “levelling out”: Both Hung and Chu (2014) and Rybicki (2006) found that there is less internal variation in a corpus of translations than in that of their source texts, that is, translations seem to be more similar among them than non-translated texts.

4. Conclusions

To conclude, corpus-based studies of literary translation are at the intersection of corpus linguistics and digital scholarship. They should combine quantity with quality, by which I mean not only that they may greatly benefit from annotation-intensive work, including manual annotation and alignment editing, but also that the analysis of corpus data should be complemented by the investigation of the wider cultural and historical context, including the circumstances of production and reception of both source and target texts. That is, while corpora and text analysis software may contribute to uncover textual and linguistic information, findings based on corpora should be assessed and interpreted in the light of extralinguistic knowledge; for as Rommel (2004: 92) so poignantly put it, “[n]o final result, let alone an "interpretation" of a text”- including a translation – “can be obtained by computing power alone; human interpretation is indispensable to arrive at meaningful results”.


Baker, Mona (1993) “Corpus linguistics and translation studies: implications and applications”, in Mona Baker, Gill Francis and Elena Tognini-Bonelli (eds) Text and Technology: In honour of John Sinclair, Amsterdam and Philadelphia: John Benjamins, 233-250.

Baker, Mona (2000) “Towards a Methodology for Investigating the Style of a Literary Translator”, Target, 12(2): 241-266.

Bosseaux, Charlotte (2001) “A Study of the Translator’s Voice and Style in the French Translations of Virginia Woolf’s The Waves”, CTIS Occasional Papers, 1: 55–75.

Bosseaux, Charlotte (2004) “Point of View in Translation: A Corpus-based Study of French Translations of Virginia Woolf’s To the Lighthouse”, Across Languages and Cultures, 5(1): 107–122.

Bosseaux, Charlotte (2007) How Does it Feel? Point of View in Translation: The Case of Virginia Woolf into French, Amsterdam: Rodopi.

Burrows, John F. (2002a) “‘Delta’: a Measure of Stylistic Difference and a Guide to Likely Authorship," Literary and Linguistic Computing, 17(3): 267–287.

Burrows, John F. (2002b) “The Englishing of Juvenal: Computational Stylistics and Translated Texts”, Style, 36(4): 677-699.

Burrows, John F. (2006) “All the Way Through: Testing for Authorship in Different Frequency Strata”, Literary and Linguistic Computing, 22(1): 27–47.

Cohen, Patricia (2010) “Digital Keys for Unlocking the Humanities’ Riches”, The New York Times, November 16, 2010 <http://www.nytimes.com/2010/11/17/arts/17digital.html?pagewanted=all> [accessed on 31/02/2015].

Frankenberg-Garcia, Ana and Diana Santos (2003) “Introducing COMPARA: the Portuguese-English Parallel Corpus”, in Federico Zanettin, Silvia Bernardini and Dominic Stewart (eds.) Corpora in Translator Education, Manchester: St Jerome, 71-87.

Gellerstam, Martin (1986) “Translationese in Swedish novels translated from English”, in Lars Wollin and Hand Lindquist (eds) Translation Studies in Scandinavia, Lund: CWK Gleerup, 88-95.

Gellerstam, Martin (2005) “Fingerprints in Translation” in Gunilla Anderman and Margaret Rogers (eds) In and Out of English: For Better, For Worse?, Clavedon: Multilingual Matters, 201-13.

Hewson, Lance (2011) An Approach to Translation Criticism. Emma and Madame Bovary in translation, Amsterdam and Philadelphia: John Benjamins.

Hoover, David L. (2008) “Quantitative Analysis and Literary Studies”, in Susan Schreibman and Ray Siemens (eds) A Companion to Digital Literary Studies, Oxford: Blackwell. <http://www.digitalhumanities.org/companionDLS/> [accessed on 31/02/2015].

Huang, Labo and Chiyu Chu (2014) “Translator’s style or translational style? A corpus-based study of style in translated Chinese novels”, Asia Pacific Translation and Intercultural Studies, 1(2): 122–141.

Hung, Jen-Jou, Marcus Bingenheimer and Simon Wiles (2010) “Quantitative evidence for a hypothesis regarding the attribution of early Buddhist translations”, Literary and Linguistic Computing, 25(1): 119-134.

Ji, Meng (2012) “Hypothesis testing in corpus-based literary translation studies”, in Michael Oakes and Meng Ji (eds) Quantitative Methods in Corpus-Based Translation Studies, Amsterdam and Philadelphia: John Benjamins, 53-72.

Ji, Meng and Michael P. Oakes (2012) “A Corpus study of early English translations of Cao Xueqin’s Hongloumeng”, in Michael Oakes and Meng Ji (eds) Quantitative Methods in Corpus-Based Translation Studies, Amsterdam and Philadelphia: John Benjamins, 175-208.

Johansson, Stig (2003) "Reflections on Corpora and their Uses in Cross-linguistic Research", in Federico Zanettin, Silvia Bernardini and Dominic Stewart (eds) Corpora in Translator Education, Manchester: St. Jerome, 135-48.

Kenny, Dorothy (2000) Lexis and Creativity in Translation: A Corpus-based Study. Manchester: St. Jerome Publishing.

Laviosa, Sara (1997) “How Comparable Can ‘Comparable Corpora’ Be?”, Target, 9(2): 289-319.

Laviosa, Sara (1998) “Core patterns of lexical use in a comparable Corpus of English narrative prose”, Meta, 43(4): 557-570.

Liu, Zequan, Liu Chaopeng, and Zhu Hong (2011) “Hong Lou Meng si ge yingyiben yizhe fengge chutan. [An exploration of the translators’ styles of four English versions of Hong Lou Meng.]”, Chinese Translators Journal, 32 (1): 60–64.

Maczewski, Jan-Mirko (1996) “Virginia Woolf's The waves in French and German waters: a computer assisted study in literary translation”, Literary and Linguistic Computing, 11(4): 175-186.

Mahlberg, Michaela (2007) “Corpora and translation studies: textual functions of lexis in Bleak House and in a translation of the novel into German”, in Vittoria Intonti, Graziella Todisco and Maristella Gatto (eds) La Traduzione. Lo Stato dell'Arte. Translation. The State of the Art, Ravenna: Longo, 115-135.

Mikhailov, Mikhail (2002) “Two Approaches to Automated Text Aligning of Parallel Texts in Fiction”, Across Languages and Cultures, 2(1): 87-96.

Munday, Jeremy (1998) “A Computer-assisted Approach to the Analysis of Translation Shifts”, Meta, 43(4): 542-556.

Olohan, Maeve and Mona Baker (2000) “Reporting that in Translated English: Evidence for Subconscious Processes of Explicitation?”, Across Languages and Cultures, 1(2): 141-158.

Øverås, Linn (1998) “In Search of the Third Code: An Investigation of Norms in Literary Translation”, Meta,43(4), 571-88. <http://id.erudit.org/iderudit/003775ar>.

Paloposki, Outi (2012) “Translation Criticism”, in Yves Gambier and Luc van Doorslaer (eds) Handbook of Translation Studies, Vol. 3, 184-190.

Price, Kenneth M. (2008) “Electronic Scholarly Editions”, in Susan Schreibman and Ray Siemens (eds) A Companion to Digital Literary Studies, Oxford: Blackwell. <http://www.digitalhumanities.org/companionDLS/> [accessed on 31/02/2015].

Rommel, Thomas (2004) “Literary Studies”, in Susan Schreibman, Ray Siemens and John Unsworth (eds) A Companion to Digital Humanities, Oxford: Blackwell Publishing, 88-96.

Rybicki, Jan (2006) “Burrowing into Translation: Character Idiolects in Henryk Sienkiewicz's Trilogy and its Two English Translations”, Literary and Linguistic Computing, 21(1): 91-103.

Rybicki, Jan (2012) “The great mystery of the (almost) invisible translator. Stylometry in translation”, in Michael Oakes and Meng Ji (eds) Quantitative Methods in Corpus-Based Translation Studies, Amsterdam and Philadelphia: John Benjamins, 231-248.

Rybicki, Jan and Magda Heydel (2013) “The stylistics and stylometry of collaborative translation: Woolf’s Night and Day in Polish”, Literary and Linguistic Computing, 28(4): 708-717.

Saldanha, Gabriela (2008) "Explicitation Revisited: Bringing the Reader into the Picture", in Juliane House (ed) Beyond Intervention: Universals in Translation Processes, trans-kom 1(1): 20-35 <http://www.trans-kom.eu/bd01nr01/trans-kom_01_01_03_Saldanha_Explicitation.20080707.pdf> [accessed on 31/02/2015].

Saldanha, Gabriela (2011a) ‘Style of Translation: The Use of Source Language Words in Translations by Margaret Jull Costa and Peter Bush’ in Alet Kruger, Kim Wallmach and Jeremy Munday (eds) Corpus Based Translation Studies: Research and Applications, London: Continuum, 237‐258.

Saldanha, Gabriela (2011b) “Translator style: Methodological considerations”, The Translator, 17(1): 25–50.

Toury, Gideon (2004) “Probabilistic explanations in translation studies: Welcome as they are, would they qualify as universals?”, in Anna Mauranen and Pekka Kujamäki (eds) Translation Universals. Do they exist?, Amsterdam and Philadelphia: John Benjamins, 15-32.

van Dalen-Oskam, Karina (2013) “Names in novels: An experiment in computational stylistics”, Literary and Linguistic Computing, 28(2): 359-370.

Wang, Qing and Li Defeng (2012) “Looking for translator's fingerprints: a corpus-based study on Chinese translations of Ulysses”, Literary and Linguistic Computing, 27(2): 81-93.

Winters, Marion (2004) “German Translations of F. Scott Fitzgerald’s The Beautiful and Damned: A Corpus-based Study of Modal Particles as Features of Translators’ Style”, in Ian Kemble (ed) Using Corpora and Databases in Translation, Portsmouth: University of Portsmouth, 71–89.

Winters, Marion (2007) “F. Scott Fitzgerald’s Die Schönen und Verdammten: A Corpus-based Study of Speech-act Report Verbs as a Feature of translator’s Style”, Meta, 52 (3): 412–425.

Winters, Marion (2009) “Modal Particles Explained: How Modal Particles Creep into Translations and Reveal Translators’ Styles”, Target, 21 (1): 74–97.

Zanettin, Federico (2000) “Parallel corpora in translation studies. Issues in corpus design and analysis”, in Maeve Olohan (ed) Intercultural Fault Lines, Manchester: St Jerome, 105–118.

Zanettin, Federico (2001) IperGrimus. inTRAlinea Monographs, <http://www.intralinea.org/monographs/zanettin/> [accessed on 31/02/2015].

Zanettin, Federico (2012) Translation-Driven Corpora. Corpus Resources in Descriptive and Applied Translation Studies, London and New York: Routledge.

Zubillaga, Zuriñe, Naroa Uribarri and Ibon Sanz (2015) “Building a trilingual parallel corpus to analyse literary translations from German into Basque”, in Claudio Fantinuoli and Federico Zanettin (eds) New Directions in Corpus-based Translation Studies, Berlin: Language Science Press. 


[1] “Burrows begins by recording the frequencies of the most frequent words of a primary set of texts by the possible authors and calculating the mean frequency and standard deviation for each word in this set of texts. He then uses z-scores to compare the difference between the mean and each of the primary authors with the difference between the mean and the questioned text for each of the words. He completes the calculation by averaging the absolute values of the z-scores of all the words to produce Delta, a measure of the difference between the test text and each primary-set author. The primary set author with the smallest Delta is suggested as the author of the test text.”(Hoover 2008: online).

[2] The results cumulated those from their study, Baker’s (2000) study, and those from yet another study, by Liu et al. (2011), who compared two different English translations of the Chinese classic novel Hongloumeng by Cao Xuequin.

[3] Saldanha (2011: 35) explains that instances of italics “were classified according to function (italics used for emphasis, italics highlighting foreign words, titles of books, words mentioned rather than used, etc.) and [according] to whether they were carried across from the source to the target text or otherwise omitted or added in the target text”.

About the author(s)

Federico Zanettin (PhD in Translation Science, University of Bologna) is Associate Professor in English Language and Translation at the University of Perugia.

Email: [please login or register to view author's email address]

©inTRAlinea & Federico Zanettin (2017).
"Issues in Computer-Assisted Literary Translation Studies"
inTRAlinea Special Issue: Corpora and Literary Translation
Edited by: Titika Dimitroulia and Dionysis Goutsos
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2254

Film Studies and Translation Studies

A Necessary Mutual Understanding

By Juan José Martínez Sierra & Beatriz Cerezo Merchan (Universitat de València, Spain)

©inTRAlinea & Juan José Martínez Sierra & Beatriz Cerezo Merchan (2017).
"Film Studies and Translation Studies A Necessary Mutual Understanding"
inTRAlinea Special Issue: Building Bridges between Film Studies and Translation Studies
Edited by: Juan José Martínez Sierra & Beatriz Cerezo Merchán
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2253

In the past two decades, the field of audiovisual translation, which is part of the Translation Studies discipline, has gained increasing interest from researchers. In fact, it is an academic field that has experienced a rapid development in a certainly short period of time. Nowadays the publications —articles and books— devoted to this field are countless. Similarly, numerous doctoral dissertations are presented every year, not to mention the growing number of international conferences that address audiovisual translation not tangentially, but exclusively. On the other hand, the origin of Film Studies dates back to almost a century ago. Their main purpose is to explore films via theoretical, historical and critical approaches.

Still, it is possible —and also necessary, as Chaume (2004) already pointed out— to continue promoting new interdisciplinary/cross-disciplinary avenues of research that lead to new breakthroughs in the two aforementioned fields. As he expresses, “[a]udiovisual texts are usually built according to the conventions of film language, a complex language that overcomes linguistic communication and has its own rules and conventions” (2004: 12). These are multimodal, multisemiotic texts that transmit information simultaneously through two different channels —acoustic and visual— and “several signifying codes which complement and frame words and linguistic meaning” (2004: 12). The links between audiovisual texts and film language are undeniable. Therefore, it is absolutely necessary to bring together the fields of Translation Studies and Film Studies, so a deeper understanding of those connections can be reached. This special issue hopes to promote an innovative perspective, in which Translation Studies and Film Studies come together.

However, although some researchers have attempted to tackle this issue (such as Chaume, 2004; Remael, 2004 and 2008; Cattrysse and Gambier, 2008; Martínez-Sierra, 2012; and some articles from the 2012 special issue of MonTI edited by Agost, Orero and di Giovanni), work of this nature is scarce, especially from a sheer cinematographic standpoint. Thus, it seems both relevant and necessary to promote further insights into the relationship between Translation Studies and Film Studies.

Once we agree that a bridge between Translation Studies and Film Studies can be built and that it creates a whole new array of research possibilities, we may foresee multiple topics to be explored —some of which are tackled in this special issue— such as audiovisual translation and film history; the influence of script writing on audiovisual translation; the influence of the technical aspects of film (i.e. camera angles, shots, camera movements…) on audiovisual translation; the influence of the signifying codes of film (i.e. linguistic, paralinguistic, musical, special effects, photographic or iconographic codes) on audiovisual translation; the film industry and audiovisual translation; technological changes on the film industry and their influence on audiovisual translation; the application of film studies to the development of creative subtitling strategies; experimental research and reception studies (that is, the impact of audiovisual translation on the audiovisual product); the inclusion of film language concepts (script writing, signifying codes, technical aspects) in audiovisual translation teaching and practice; accessible filmmaking and the integration of audiovisual translation and accessibility during the filmmaking process; remakes as a form of translation (the creation of new audiovisual texts based on the adaptation of foreign products); translating from the page to the screen (the adaptation of comics into films); cinema, identity and translation; cinema, ideology and translation; or cinema, politics and translation.

This special issue gathers a series of contributions, both theoretical and applied, that bridge the gap between these two disciplinary areas and explore the links between them. Academics and professionals from the fields of Translation and Film Studies have contributed to it with articles covering a wide range of audiovisual translation modes (dubbing, interlingual subtitling, subtitling for the deaf and hard-of-hearing, audio description, video games, etc.), film genres (films, cartoons, documentaries, video games, etc.) and audiences (adults and children with or without special needs). This variety is intentionally motivated by the desire and the need to offer a panoramic view of the research that is being carried out on the intersection between audiovisual translation and Film Studies. All articles also reveal a clear interdisciplinary nature in their methodological approaches and are representative of a wide range of new avenues of research. For example, some of them use and/or propose descriptive models that combine translation aspects and cinematographic elements to analyse audiovisual products and that could be extremely useful for professionals, researchers, teachers or students of both disciplines. Similarly, some others present exploratory approaches that include alternative proposals to create audiovisual products and translations that are more creative and inclusive, and that could improve the overall experience of audiences in the present and near future.

To sum up, we consider this is an innovative volume that hopes to encourage a line of research that can certainly continue to be followed in future contributions. We need more studies that carry on bridging the gap between Translation Studies and Film Studies, and that also take into account other disciplines to help build broader and more comprehensive and interdisciplinary approaches. In addition, we should join individual and collective efforts and interests if we want to make our disciplines evolve.


Agost, Rosa, Pilar Orero and Elena di Giovanni (eds.) MonTI, vol. 4, special issue on Multidisciplinarity in Audiovisual Translation.

Cattrysse, Patrick and Yves Gambier (2008) “Screenwriting and translating screenplays” in The Didactics of Audiovisual Translation, Jorge Díaz-Cintas (ed.), Amsterdam & Philadelphia, John Benjamins: 39-55.

Chaume, Frederic (2004) “Film Studies and Translation Studies: Two Disciplines at Stake in Audiovisual Translation”, Meta: Translators' Journal, vol. 49, no. 1: 12-24.

Martínez Sierra, Juan José (2012) “On the Relevance of Script Writing Basics in Audiovisual Translation Practice and Training”, Cadernos de Tradução, vol. 29, no. 1: 145-163.

Remael, Aline (2004) “A place for film dialogue analysis in subtitling courses” in Topics in Audiovisual Translation, Pilar Orero (ed.), Amsterdam & Philadelphia, John Benjamins: 103-126.

--- (2008) “Screenwriting, scripted and unscripted language. What do subtitlers need to know?” in The Didactics of Audiovisual Translation, Jorge Díaz-Cintas (ed.),  Amsterdam & Philadelphia, John Benjamins: 57-67.

About the author(s)

Dr. Juan José Martínez Sierra works as an Associate Professor in the Department of English and German Languages and Cultures at the Universitat de València, where he teaches Written and Audiovisual Translation, Intercultural Communitation, and English Language (undergraduate and graduate). In addition to a doctorate in Translation Studies (UJI, 2004), he holds a degree in English Language and Culture (UJI, 1995) and an MA in Intercultural Communication (University of Maryland, Baltimore County, USA, 2001). He is specialized in Audiovisual Translation, a field to which he devotes both part of his teaching work and his research activity, which has mainly been focused on the study of audiovisual translation from an intercultural perspective. To date, this activity has been generously fruitful in the form of lectures, seminars, invited talks and papers at conferences, both national and international. Besides, he has published numerous works, including five books, several book chapters, reviews, and many other pieces of research in the form of articles in prestigious scientific journals. He coordinates CiTrans, and also collaborates with the research groups TRAMA (Universitat Jaume I) and SILVA (Universitat de València).

Beatriz Cerezo Merchán Holds a BA and a PhD in Translation and Interpreting. At present she is a lecturer of translation and English language at the Universitat de València. She is a member of the research groups CiTrans (http://citrans.uv.es/) and TRAMA (http://www.trama.uji.es/), and her main areas of research are the didactics of translation, and audiovisual translation and accessibility.

Email: [please login or register to view author's email address]

©inTRAlinea & Juan José Martínez Sierra & Beatriz Cerezo Merchan (2017).
"Film Studies and Translation Studies A Necessary Mutual Understanding"
inTRAlinea Special Issue: Building Bridges between Film Studies and Translation Studies
Edited by: Juan José Martínez Sierra & Beatriz Cerezo Merchán
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2253

Montaje audiovisual e integración de la audiodescripción en la producción documental

Estudio de caso del documental Vidas Olímpicas

By B. Cerezo Merchán, I. de Higes Andino, E. Galán, & R. Arnau Roselló (Universitat de València & Universitat Jaume I, Spain)

Abstract & Keywords


This article is the fruit of a multidisciplinary research project that links the fields of audiovisual translation and audiovisual communication: the project ITACA (Social Inclusion, Audiovisual Translation and Audiovisual Communication), that revolves around the concept of accessible filmmaking (Romero Fresco 2013). This project suggests considering the needs of people with sensory disabilities during the processes of audiovisual pre-production, production and post-production to improve the level of accessibility of audiovisual works. The present study focuses on discovering if editing techniques could be assembled in a different way in an audiovisual work, in order to facilitate its audio description. Additionally, investigation is done on how the editing sequence could be used during the editing process as a tool by the audiovisual professional to detect those excerpts that would be easier to audio describe and those that would prevent a correct audio description due to the absence of gaps in the audio track. This article is based on the analysis of an accessible translation mode (audio description) in a specific audiovisual genre (the documentary). More specifically, we analyse the documentary Vidas Olímpicas (César Martí 2015), how it was originally edited and how the editing could be performed differently to facilitate the addition of audio description. Firstly, this article provides a theoretical introduction to audiovisual accessibility and to the concept of accessible filmmaking. Secondly, the objectives and hypotheses of the study, as well as the documentary Vidas Olímpicas, are presented. Next, the working methodology and the analysis, carried out in collaboration by experts in audiovisual translation and audiovisual communication, are explained. Finally, conclusions are drawn from the results of the analysis and change proposals in the productive routines of scriptwriting, filming and editing of the audiovisual work – which might ultimately ease the process of audio description – are offered.


Este artículo es fruto de un proyecto de investigación multidisciplinar que une los ámbitos de la traducción audiovisual y la comunicación audiovisual: el proyecto ITACA (Inclusión social, Traducción audiovisual y Comunicación audiovisual), que gira en torno al concepto de producción audiovisual accesible o accessible filmmaking (Romero Fresco 2013). Este proyecto propone tener en cuenta las necesidades de las personas con discapacidad sensorial durante los procesos de preproducción, producción y postproducción de las obras audiovisuales para mejorar su nivel de accesibilidad. El presente trabajo se centra en descubrir si existen elementos del montaje audiovisual de una obra que podrían editarse de forma diferente para facilitar su audiodescripción y, por tanto, mejorar su accesibilidad. Asimismo, se estudia cómo podría emplearse la secuencia de montaje como herramienta que permita al montador profesional identificar, durante el proceso de montaje, qué fragmentos son más fácilmente audiodescriptibles y cuáles menos, debido a la inexistencia de huecos de mensaje. Este artículo se basa en el análisis de una modalidad de traducción accesible (la audiodescripción) en un género audiovisual (el documental). Concretamente, se analiza el documental Vidas Olímpicas (César Martí 2015), cómo se editó originalmente y cómo se podría modificar su edición para facilitar la inclusión de audiodescripción. En primer lugar, en este artículo se realiza una introducción teórica sobre accesibilidad audiovisual y sobre el concepto de producción audiovisual accesible. En segundo lugar, se presentan los objetivos y las hipótesis del estudio, así como el documental Vidas Olímpicas. A continuación, se exponen la metodología de trabajo y el análisis, realizado en colaboración por expertos en traducción audiovisual y comunicación audiovisual. Por último, se extraen conclusiones a partir de los resultados del análisis y se presentan propuestas de cambio en las rutinas productivas de guion, rodaje y montaje de la obra audiovisual que, en última instancia, podrían facilitar el proceso de audiodescripción.

Keywords: audio description, documentary, editing, audiodescripción, documental, producción audiovisual accesible, audiovsual production, accessible filmmaking, creación audiovisual, montaje audiovisual

©inTRAlinea & B. Cerezo Merchán, I. de Higes Andino, E. Galán, & R. Arnau Roselló (2017).
"Montaje audiovisual e integración de la audiodescripción en la producción documental"
inTRAlinea Special Issue: Building Bridges between Film Studies and Translation Studies
Edited by: Juan José Martínez Sierra & Beatriz Cerezo Merchán
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2252

1. Introducción

Durante muchos años, investigar en traducción audiovisual (TAV) significaba abordar las dos modalidades más extendidas globalmente, es decir, el doblaje y la subtitulación, así como otras modalidades menores como las voces superpuestas (voice-over), la interpretación simultánea, la narración, el doblaje parcial, el comentario libre o la traducción a la vista (Chaume 2004: 31-40). Sin embargo, a finales del siglo XX, gracias a los esfuerzos de nuestra sociedad para que toda la información esté al alcance de todos los ciudadanos y a la constante evolución de las nuevas tecnologías, se afianza la traducción para la accesibilidad. Las dos modalidades principales de TAV que se emplean a día de hoy para hacer accesibles los productos audiovisuales a personas con discapacidades sensoriales son la subtitulación para personas sordas o con discapacidad auditiva (SPS)[1] y la audiodescripción (AD) para personas ciegas o con discapacidad visual[2]. También es posible hacer accesibles los productos audiovisuales mediante la interpretación en lengua de signos o, incluso, la audiosubtitulación, las cuales quedan fuera del alcance de este trabajo.

La primera de estas dos modalidades accesibles, la SPS, se define como el medio de apoyo que presenta un ‘conjunto de texto y gráficos que muestran en pantalla los discursos orales, la información contextual y los efectos sonoros que se producen en cualquier obra audiovisual’ (AENOR 2012: 7). Así, este tipo de subtitulación se caracteriza por emplear colores u otras técnicas para identificar a los personajes, así como por añadir información que especifica el modo en el que se articulan los diálogos, los sonidos que se escuchan o la música que acompaña las acciones.

Por su parte, la AD se describe como un servicio de apoyo a la comunicación que consiste en un ‘conjunto de técnicas y habilidades aplicadas, con objeto de compensar la carencia de captación de la parte visual contenida en cualquier tipo de mensaje, suministrando una adecuada información sonora que la traduce o explica’ (AENOR 2005: 4). De esta forma, el mensaje llegaría a percibirse como un todo armónico y de la forma más parecida posible a como lo recibiría una persona sin discapacidad visual. Concretamente, la AD permite narrar la acción y describir la ambientación, el lugar en el que transcurre la acción, los sonidos que no son fácilmente distinguibles, así como leer cualquier tipo de texto que se vea en pantalla. Para ello se aprovechan los llamados ‘huecos de mensaje’, es decir, aquellos espacios que suelen estar ‘limpio[s] de diálogos y de sonidos relevantes para la comprensión de la obra’ (AENOR 2005: 5) y que son necesarios para poder incluir cada uno de los ‘bocadillos de información’ o ‘unidades descriptivas’.

Al objeto de establecer y asegurar un acceso igualitario a los medios audiovisuales, gobiernos de distintos países han desarrollado normativas que regulan los porcentajes mínimos de emisiones televisivas que deben hacerse accesibles. Estas normativas nacionales[3] y la futura European Accessibility Act, cuya propuesta se presentó en diciembre de 2015[4], están fomentando un considerable aumento del número de productos audiovisuales accesibles en muchos países como Francia, Reino Unido, Estados Unidos, Australia y España. En el caso de España, la Ley 7/2010, de 31 de marzo, General de la Comunicación Audiovisual[5] ha impulsado el crecimiento del mercado profesional de la accesibilidad desde su aprobación, ya que determina un aumento progresivo del volumen de emisiones accesibles anuales hasta 2013 (en el caso de la SPS, con el fin de alcanzar el 90 por ciento de emisiones subtituladas en las cadenas públicas y el 75 por ciento en las privadas, y en el caso de la AD, diez horas a la semana en cadenas públicas y dos horas en las cadenas privadas)[6]. Queda fuera de esta regulación la programación audiovisual disponible a través de otros medios, como los servicios de televisión a la carta y las plataformas de visionado por Internet (CNMC 2016). Cabe destacar la incorporación de SPS en algunos servicios de televisión a la carta (por ejemplo, de RTVE y TV3), así como distintas aplicaciones para smartphones y tablets, que ofrecen AD, SPS y/o lengua de signos (por ejemplo, la app 5S para los contenidos de Movistar+[7], la app AudescMobile que lanzó la ONCE y Fundación Vodafone[8] y la app WhatsCine, creada por esta empresa con la colaboración de la Universidad Carlos III de Madrid[9]).

Asimismo, de manera que este incremento de emisiones accesibles para los colectivos con discapacidades sensoriales vaya acompañado de una mejora de su calidad, cada vez son más los países que están trabajando en la elaboración de estándares de calidad de las prácticas de traducción audiovisual accesible. España, por ejemplo, cuenta con la norma UNE 153010 (2012), que regula la práctica de la SPS y actualiza la norma previa de 2003, y la norma UNE 153020 (2005), que regula la práctica de la AD[10].

El mercado laboral de la traducción audiovisual, en general, y de la traducción para la accesibilidad, en particular, está en auge en cuanto al volumen (CNMC 2016). Además, su interés científico y académico es también cada vez mayor. Ha proliferado la publicación de libros, artículos y tesis (Matamala y Orero 2009 y 2016; Arnáiz Urquiza 2012; Cabeza Cáceres 2013; Fresno Cañada 2014; Maszerowska, Matamala y Orero 2014; Romero Fresco 2015; Tamayo Masero 2015; Fryer 2016; Muller 2016, etc.); la organización de congresos y conferencias internacionales (como por ejemplo, el AMADIS[11] desde 2006 y el ARSAD[12] desde 2007); la incorporación de asignaturas o itinerarios de especialidad en TAV y en accesibilidad en estudios de grado y postgrado, etc. (Cerezo Merchán y de Higes Andino 2013).

Pese a ello, esta disciplina no ha logrado aún superar un importante obstáculo: su invisibilidad en los Estudios Fílmicos o la industria audiovisual[13]. Estudios como los de Lambourne (2012) o Romero Fresco (2013) denuncian el desequilibrio existente entre la inversión que la industria cinematográfica estadounidense realiza en servicios de traducción audiovisual y de accesibilidad (entre un 0,1 y un 1 por ciento del presupuesto medio de cualquier producción cinematográfica) y los beneficios que sus producciones reportan. Con el propósito de integrar la traducción y la accesibilidad como parte del proceso de producción audiovisual adoptamos, en el siguiente epígrafe, el concepto de producción audiovisual accesible, en torno al que gira el proyecto de investigación del que se deriva este artículo.

A continuación, se presenta, en el apartado 2, el concepto teórico central de nuestro estudio: la producción audiovisual accesible. Tras ello, se describen los objetivos, las hipótesis de partida, la metodología y el corpus de nuestra investigación en el apartado 3. Por último, se analizan los datos recabados en el apartado 4 y se establecen conclusiones en el apartado 5.

2. La producción audiovisual accesible

El proyecto de investigación del que hablamos tiene por título ‘Inclusión social, Traducción audiovisual y Comunicación audiovisual’ (ITACA) y cuenta con la financiación del Ministerio de Economía y Competitividad de España. Con referencia FFI2016-76054-P y adscrito a la Universitat Jaume I, se trata de un proyecto en el que, desde 2017, trabajan conjuntamente académicos y profesionales de los ámbitos de la traducción y la comunicación audiovisual con el objetivo de establecer pautas mediante las que se integre la accesibilidad en las fases de preproducción, producción y postproducción, y difusión de cualquier género o formato audiovisual.

Como se avanza en la introducción, este proyecto de investigación adopta el concepto de producción audiovisual accesible o accessible filmmaking, que se fundamenta en la integración de la traducción y la accesibilidad como parte del proceso de producción audiovisual y está empezando a desarrollarse en el Reino Unido[14], sobre todo en películas independientes (véase Romero Fresco 2013).

Incorporar el concepto de producción audiovisual accesible en la preproducción y la producción audiovisual se encuentra en línea con el ‘diseño para todos’ (Mace 1988) o ‘diseño universal’ (ONU 2006), que cada vez cobra más fuerza en nuestras sociedades refrendado por leyes de igualdad de oportunidades, no discriminación y accesibilidad. El diseño para todos o universal consiste en el diseño de productos, entornos y servicios aptos para el uso del mayor número de personas sin necesidad de adaptaciones posteriores ni de un diseño especializado.

Actualmente las actividades que posibilitan la accesibilidad de los productos audiovisuales para personas con discapacidad sensorial, es decir, la subtitulación para personas sordas y la audiodescripción, no cumplen los requisitos del diseño para todos, ya que no se tienen en cuenta desde el inicio de la realización audiovisual (Udo y Fels 2009). El hecho de que no se tengan en cuenta determinados criterios para hacer los productos audiovisuales más fácilmente accesibles desde el inicio dificulta posteriormente la tarea de subtitulación y audiodescripción e, incluso, puede hacerla total o parcialmente imposible, como muestra el hecho de que en España prácticamente solo se subtitule y audiodescriba ficción (CERMI 2015). Este último dato podría sugerir que, por una parte, existen limitaciones en los rasgos de género textual que caracterizan los formatos televisivos y que estarían dificultando a las cadenas hacer accesible parte de su parrilla de programación[15]. Por otra parte, también es cierto que algunos formatos televisivos son accesibles para las personas con discapacidad visual sin necesidad de ser audiodescritos. Por ejemplo, los formatos televisivos que se emiten en horario de máxima audiencia (prime time), como los informativos diarios, ya que estos son concebidos para un público que, en muchos casos, tiene la atención dividida entre la pantalla de televisión, otras pantallas o tareas cotidianas (Fernández Fernández y Galán 2011).

A continuación, se presentan los objetivos e hipótesis de partida, el corpus y la metodología de análisis de este estudio, que forma parte, como se ha expuesto anteriormente, de un proyecto de investigación de mayor envergadura: el proyecto ITACA.

3. Descripción de la investigación

El género audiovisual elegido para su análisis es el documental, concepto al que otorgaremos un sentido puramente semántico, sin entrar en cuestiones de fondo que atañen, entre otros, a la distinción entre documental y reportaje o a los límites de la representación de la realidad, dado que no es ese el objetivo específico de esta investigación. Asimismo, la modalidad accesible analizada es la audiodescripción.

3.1. Objetivos e hipótesis de partida

Nuestro objetivo principal en este trabajo es analizar cómo integrar la audiodescripción en el montaje audiovisual, es decir, detectar aquellos elementos que pueden grabarse de otro modo para que el montaje final facilite la audiodescripción. De este objetivo principal emanan tres objetivos específicos. Por una parte, estudiar de qué forma se puede utilizar la secuencia de montaje como herramienta que ayude al profesional del audiovisual a identificar más fácilmente los huecos de mensaje. Por otro lado, demostrar la idoneidad de trabajar en un equipo interdisciplinar compuesto por expertos en traducción audiovisual y comunicación audiovisual y, por último, adoptar una terminología común para hacer referencia a la audiodescripción y a los elementos y los procesos propios de una obra audiovisual. De este modo, el diálogo es fluido entre estos expertos, lo que puede ser relevante de cara a su implantación futura en la práctica habitual en los procesos de producción y traducción audiovisual.

En esta investigación, partimos de la hipótesis de que la utilización de la secuencia de montaje permitirá al profesional del audiovisual prever durante el proceso de montaje aquellos fragmentos más fácilmente audiodescriptibles e, incluso, aquellos que impedirían una correcta audiodescripción por la inexistencia de huecos. Con ello, se garantizaría una mayor calidad de los productos audiovisuales, que desde un primer momento serían concebidos y diseñados teniendo en cuenta las necesidades del amplio abanico de público al que se dirigen, con o sin discapacidades visuales, y por lo tanto, teniendo también en cuenta las tareas de audiodescripción.

3.2. Corpus de análisis

El documental seleccionado para este análisis se titula Vidas Olímpicas (César Martí 2015)[16]. Se trata de un proyecto promovido por la Fundación Trinidad Alfonso y dirigido por César Martí que fue estrenado en noviembre de 2015 en el evento Impuls-a València y emitido ese mismo mes en Teledeporte (canal especializado en deportes de la Corporación de Radio y Televisión Española y, por tanto, público y de emisión nacional).

Para poder realizar este estudio se necesitaba la implicación del equipo de producción de la obra audiovisual analizada y, a tal efecto, se consiguió la colaboración del director César Martí, que nos dio acceso (y permiso para su análisis y reproducción) a las siguientes fuentes de material original:

  • Una copia del documental Vidas Olímpicas
  • El documento de pitching[17]
  • La escaleta donde se describen las líneas narrativas del documental
  • Capturas de pantalla del proyecto de montaje del documental

En el documento de pitching se explica que: ‘no es un documental de deportes, es un documental de cinco historias humanas de superación y esfuerzo. Es un documental que nos habla de la vida, de las alegrías y decepciones que nos hacen crecer como persona y de un conjunto de deportistas [olímpicos y paralímpicos] que se superan día a día en su camino hacia las Olimpiadas Río 2016’. En el documental se intercalan las historias de los cinco deportistas protagonistas. Para conocerlos más a fondo, sus intervenciones se combinan con imágenes de su día a día y con entrevistas a familiares, parejas y compañeros de profesión. Asimismo, otras grandes figuras del deporte hablan también de estos deportistas y de los valores del olimpismo desde su perspectiva.

La selección de la muestra que se analiza ha respondido a un criterio compartido de síntesis y representatividad. Dado que el artículo explora las posibilidades de incorporar al proceso de producción documental algunas pautas para facilitar la integración de la audiodescripción en el audiovisual, el análisis y la ejemplificación de la propuesta a través de un solo documental nos proporciona un marco de referencia que no altera en esencia sus características. Por ello, es un elemento que permite la identificación de las estrategias que se incluyen en este artículo en un texto más estable desde el punto de vista de la forma y la narración. Hemos seleccionado tres fragmentos para su análisis, entre los que se encuentran dos del nadador paralímpico Ricardo Ten y uno del equipo de gimnasia:

  • 00.00.00-00.05.56 Ricardo Ten (inicio del documental y presentación del personaje)
  • 00.17.36-00.22.43 Ricardo Ten (tratamiento del personaje en competición)
  • 00.57.17-01.02.37 (clímax final del equipo de gimnasia con montaje paralelo entre ensayo y competición y fin del documental)

La selección de estos tres fragmentos se ha visto determinada por su capacidad metonímica, es decir, por su representatividad como parte de un todo más amplio que sería, en este caso, el documental Vidas Olímpicas completo[18]. Así, cada uno de los fragmentos analizados se refiere a un estadio narrativo diferente en el que el estatuto del narrador se adapta a las acciones mostradas, de tal modo que el análisis de los tres fragmentos nos permite comprender cómo proceder en estos contextos concretos. Por consiguiente, recogemos el inicio y el final del documental y un fragmento de competición. Además, como dato digno de reseñar, llama la atención que no varía de forma alguna el tratamiento visual que se da a los deportistas olímpicos y a los deportistas paralímpicos[19].

En este trabajo partimos de la premisa de que el montaje audiovisual se realiza en la fase de postproducción, pero que en realidad durante la preproducción se realiza una planificación previa que contiene ‘la selección y combinación de imágenes y sonidos destinados a la construcción del significante’ (Fernández Díez y Martínez Abadía 1999: 174). Dado que se tiene en cuenta dicha planificación durante el rodaje, el objetivo de este análisis es detectar aquellos elementos que pueden grabarse de otro modo para que el montaje final facilite la audiodescripción.

3.3. Fases de análisis

En los siguientes apartados se lista la terminología acordada por los autores y se describen las dos fases de análisis llevadas a cabo a partir de la postproducción de este documental: el análisis descriptivo de la planificación audiovisual y la creación de una propuesta alternativa de secuencia de montaje.

3.3.1. Definiciones de terminología empleada para describir las técnicas narrativas

Como resultado del objetivo que persigue adoptar una terminología común para hacer referencia a los elementos y los procesos propios de una obra audiovisual, de modo que el diálogo sea fluido entre estos expertos, en este subapartado se definen las técnicas narrativas empleadas en el documental analizado según la terminología clásica procedente del análisis fílmico (Marzal-Felici y Gómez-Tarín 2015; Galán 2008).

  • Cartón de grafismo. Imagen a total de procedencia infográfica que tiene algún tipo de valor informativo.
  • Edición y montaje audiovisual. ‘La selección y combinación de imágenes y sonidos destinados a la construcción del significante’ (Fernández Díez y Martínez Abadía 1999: 174). Aunque es habitual que se utilicen como sinónimos, la edición se refiere habitualmente al proceso mecánico de unir planos e introducir transiciones, mientras que el montaje alude más al proceso intelectual de dar sentido a una historia a través de la sucesión de imágenes.
  • Encabalgado de audio o audio split. El sonido se adelanta a la aparición de la fuente que lo produce en imagen de modo que no se perciben conjuntamente en el mismo momento y se suspende por unos segundos el sentido de la narración. Se intenta que el audio y el vídeo no tengan siempre una coincidencia total para estimular la pulsión escópica del espectador. Normalmente las fuentes de imagen relevantes se comienzan a escuchar siempre antes de verlas (así incentivamos el deseo de saber del espectador).
  • Flashback (o analepsis). Elemento expresivo de la narración por cuya presencia la acción se traslada a un momento del pasado para después recuperar el momento presente de la historia. Permite recuperar conceptos o personajes y resumir un contenido o reafirmar una tesis.
  • Fuera de campo. Espacio fílmico sugerido por el campo visual que incluye todo aquello que no se haya representado icónicamente en el plano y queda fuera del cuadro.
  • Montaje paralelo o alterno. Método de montaje que consiste en la intercalación sucesiva de planos relativos a dos o más acciones simultáneas o no. Permite también realizar asociaciones conceptuales o conectar determinadas secuencias del film.
  • Plano recurso. Es todo aquel plano que se toma en un momento determinado del rodaje o se recupera del archivo para que pueda ser incorporado en el montaje final sin que se encuentre determinado el lugar exacto en el que se va a incorporar. En las piezas informativas y de carácter documental fundamentalmente hay dos tipos de plano: las entrevistas a total y los planos recurso.
  • Plano. Espacio entre dos cortes de montaje. Se trata de la unidad mínima narrativa en los textos audiovisuales.
  • Pulsión escópica. Es un término que hace referencia al impulso de mirar, así como al deseo de ser mirado. Este concepto esconde muchas de las claves del montaje cinematográfico (Gómez López 2013).
  • Rótulo. Sobreimpresión en una parte de la pantalla que aporta algún tipo de información que identifica, contextualiza o ubica el contenido de la imagen.
  • Sonido ambiente. Conjunto de sonidos propios del espacio dramático y de la acción que se escuchan en pantalla. Pueden ser grabados durante la filmación o añadidos en la postproducción. Cuando estos sonidos contienen, además, información de carácter verbal que acompaña a la acción representada, hablamos de sonido ambiente significativo.
  • Texto en pantalla. Se refiere a los caracteres alfanuméricos que ofrecen información sobre los personajes y su ubicación espacio-temporal. El texto en pantalla también conocido como rótulos también se emplean para subtitular una intervención o para introducir información complementaria al discurso.
  • Títulos de crédito. Son aquellas partes del relato audiovisual, generalmente construidos a partir de rotulación, que informan al espectador del título y de la autoría de la obra audiovisual.
  • Total. Declaración a cámara de un personaje que aparece en imagen, que generalmente forma parte de una entrevista.
  • Voz en off. Aquella que proviene de un personaje que no se encuentra en el interior del campo visual pero sí forma parte de la diégesis.

3.3.2. Análisis descriptivo de la planificación audiovisual

En esta primera fase de análisis, se elabora una descripción de la planificación audiovisual de forma conjunta entre los expertos en traducción audiovisual y comunicación audiovisual, autores de este trabajo.

Para ello, en primer lugar, las especialistas en traducción audiovisual y accesibilidad desglosan la planificación audiovisual atendiendo a los aspectos que hay que audiodescribir para el colectivo con problemas de visión y a los cortes de montaje, sin dar un carácter diferenciado a los tipos de plano (no se distingue, por ejemplo, entre plano master y planos de detalle salvo que su distinción resulte relevante para el receptor). En segundo lugar, los especialistas en comunicación audiovisual describen las técnicas narrativas que se perciben en las muestras (v. apartado 3.3.1. para una definición de todas las técnicas narrativas). En tercer lugar, se reflexiona sobre el elemento visual (imagen o texto en pantalla) o sonoro que convendría audiodescribir para su comprensión por un público con problemas de visión. En cuarto lugar, se concluye si la pista de diálogos permite inferir los elementos anteriormente detectados, en cuyo caso no sería necesaria la descripción para no repetir información. Por último, como la audiodescripción se debe incluir en los huecos de mensaje, se detecta si existen espacios limpios de diálogos y de sonidos relevantes, ya sea en simultaneidad con la acción o con antelación o retraso.

Toda esta información queda recogida en una hoja de Excel disponible en línea[20], en la que se analiza la siguiente información:

  • Ref. de muestra: numeración correlativa de todas las muestras de los tres fragmentos.
  • TCR: tiempos de entrada y de salida de las muestras.
  • Descripción imágenes: desglose de la planificación audiovisual.
  • Técnica narrativa: técnica o técnicas narrativas que se percibe(n) en las muestras.
  • Elemento que sería necesario describir: elemento visual (imagen o texto en pantalla) o sonoro que convendría audiodescribir para su comprensión por un público con problemas de visión.
  • Análisis del diálogo: Descripción de si la pista de diálogos permite inferir los elementos que, en general, se incluirían en una AD.
  • Hueco de mensaje: Detección de espacios limpios de diálogos y de sonidos relevantes.

3.3.3. Creación de una propuesta alternativa de secuencia de montaje

Tener acceso al material original de trabajo nos permitió constatar que la organización de la secuencia de montaje (timeline) no permitía una identificación de los huecos de mensaje ni de la tipología de planos utilizada. A continuación, presentamos una captura de pantalla de la secuencia de montaje original, donde esto se puede observar claramente:


Figura 1. Imagen de la secuencia de montaje original.
Fuente: Anlo produccions

La segunda fase del análisis ha consistido en la creación de un timeline, en el que el vídeo y el audio se han separado en tres pistas diferentes respectivamente, según muestra la siguiente figura:


Figura 2. Desglose de las pistas de vídeo y audio con la descripción del contenido


Figura 3. Reconstrucción de la secuencia de montaje del fragmento inicial
realizado con el software de edición de vídeo Adobe Premiere Pro (TCR 00.00.00-00.05.56).
Fuente: Elaboración propia a partir del documental analizado

En el caso del audio se ha discriminado el que se utiliza de fondo durante las entrevistas porque no es reseñable para la audiodescripción (es posible añadir unidades descriptivas sobre conversaciones de fondo).

Las figuras que presentamos (figuras 3, 4 y 5) muestran la reconstrucción de los tres fragmentos del documental seleccionados para su análisis con las pistas de vídeo y audio ordenadas por contenido según lo siguiente:

  • Pista 1 de vídeo y audio: entrevistas.
  • Pista 2 de vídeo y audio: planos recurso y sonido ambiente.
  • Pista 3 de vídeo y audio: cartones de grafismo y música.


Figura 4. Reconstrucción de la secuencia de montaje del segundo fragmento (TCR 00.17.36-00.22.43).
Fuente: Elaboración propia


Figura 5. Reconstrucción de la secuencia de montaje del fragmento final (TCR 00.57.17-01.02.37).
Fuente: Elaboración propia

Este desglose parece simplificar la estructura del montaje original y tal vez permita identificar a simple vista si los fragmentos pueden ser audiodescritos o no.

4. Análisis de datos

Tras el análisis de los tres fragmentos seleccionados, identificamos las siguientes seis técnicas, que son características del género documental (Marzal y Gil 2008) y que, en cierta medida, dificultan una correcta audiodescripción por la inexistencia de huecos. El análisis descriptivo de la planificación audiovisual nos permite plantear una propuesta alternativa de secuencia de montaje para que el montaje final facilite la audiodescripción. Las seis técnicas identificadas, definidas en 3.3.1, son:

  • La aparición de texto en pantalla
  • La presencia de rótulos sobre totales
  • La combinación de planos recurso y totales con rótulos
  • El encabalgado de audio sobre imágenes
  • El montaje paralelo de imágenes
  • El flashback

4.1. La aparición de texto en pantalla

Como recomiendan Matamala y Orero (2015), hay que determinar, en primer lugar, de qué tipo de texto en pantalla se trata. En este documental, tanto al principio como al final aparecen títulos de crédito en pantalla no narrados por una voz, sobre fondo negro y con música de fondo y, por tanto, texto en pantalla extradiegético.

Por una parte, tenemos cartones de grafismo con la información de autoría (copy) que presentan: el logotipo de la productora (muestras 1 y 134) y el título del documental (muestras 2 y 123) tanto al principio como al final del documental, y los logotipos del proyecto deportivo FER, impulsado por la Fundación Trinidad Alfonso, (muestra 125) y de la fundación en sí (muestra 135) solo al final de la obra.


Fotograma 1. Texto en pantalla. Logotipo de la productora (TCR 00.00.00-00.00.03 y 01.02.30-01.02.32, muestras 1 y 134)


Fotograma 2. Texto en pantalla. Título del documental (TCR 00.00.05-00.00.07 y 01.01.39-01.01.42, muestras 2 y 123)


Fotograma 3. Texto en pantalla. Logotipo del proyecto deportivo FER (TCR 01.02.00-01.02.01, muestra 125)


Fotograma 4. Texto en pantalla. Logotipo de la Fundación Trinidad Alfonso (TCR 01.02.33-01.02.36, muestra 135)

Todos ellos son ejemplos de texto en pantalla que se suele audiodescribir en cualquier obra audiovisual por ofrecer información relevante y única que se presenta solo de forma visual (el logotipo de la productora, el título del documental y los logotipos del proyecto y la fundación que lo financia). En todos estos casos hay huecos de mensaje apropiados para la audiodescripción del texto que aparece en pantalla (el texto en pantalla aparece sobre fondo negro y, por tanto, no coincide con acción que prime sobre el texto en pantalla, Matamala 2014). El audiodescriptor tendrá que decidir, como señalan Matamala y Orero (2015), si se incorpora una explicación antes de su lectura, se cambia la entonación o se utiliza otro locutor para su narración.

Por otra parte, también como texto en pantalla, al final del documental encontramos los títulos de crédito. En primer lugar, tenemos un cartón de grafismo, que contiene un crédito con información sobre el proyecto deportivo impulsado por la fundación y sobre la pertenencia de los deportistas protagonistas del documental a este (muestra 124). Por la importancia de la información presentada y por ser, de nuevo, transmitida únicamente por el canal visual, este tipo de créditos siempre se audiodescribe. En el caso de este documental, consideramos, asimismo, que el hueco de mensaje es suficiente para su audiodescripción, pues no coincide con ningún tipo de acción.


Fotograma 5. Texto en pantalla. Crédito con información sobre el proyecto FER y deportistas protagonistas del documental (TCR 01.01.43-01.01.59, muestra 124)

En segundo lugar, contamos con títulos de crédito de autoría (muestras 126 a 129) y de agradecimientos (muestras 130 a 133). Generalmente, en las obras audiovisuales solo se audiodescriben los créditos finales principales (producción, realización, guion, etc.), dejando de lado el resto del listado de créditos por considerarse secundarios, tal y como recomienda Matamala (2014). Según esta premisa, en esta obra solo se debería audiodescribir el título de crédito de autoría principal (muestra 126). Los tres segundos (01.02.02-01.02.05) en los que aparece este crédito en pantalla no bastan para audiodescribirlo, pero se podría aprovechar el hueco de mensaje de los siguientes títulos de créditos, que ya no se audiodescribirían.


Fotograma 6. Texto en pantalla. Título de crédito de autoría principal del documental (TCR 01.02.02-01.02.05, muestra 126)

Llegamos así a la conclusión preliminar de que todo texto susceptible de ser audiodescrito (en esta obra, cartones de grafismo y títulos de crédito principales) debería contar con huecos de mensaje suficientes para su lectura en voz alta (su audiodescripción). Esta conclusión coincide con la propuesta de Romero Fresco (2013) de alargar la duración de planos que contengan información textual en pantalla.

4.2. La presencia de rótulos sobre totales

Una de las técnicas narrativas frecuentes que caracteriza al género documental es el uso de totales, que son, como se definió en el apartado 3.3.1, las declaraciones a cámara de un personaje que aparece en imagen. En general, como sucede en el fotograma 7, el primer total del entrevistado viene acompañado de un rótulo en el que se especifica quién es. Por tanto, como señalan Matamala y Orero (2015), el audiodescriptor tendrá que analizar si este rótulo no diegético es redundante con la intervención del personaje, porque él mismo ofrezca esa información, o si, por el contrario, debe incluirse en la audiodescripción (ya sea de forma literal o condensada). En este caso, también será clave decidir si se puede incluir de forma sincrónica o si debe avanzarse o posponerse cuando el rótulo que identifica al personaje que está en pantalla aparezca una vez iniciada la intervención.


Fotograma 7. Total de Miki Oca en Vidas olímpicas (TCR 00.04.03, muestra 23)

Entre los TCR 00.18.55 y 00.20.24 (muestras 33 a 40), encontramos ocho totales seguidos. Cuatro de ellos, además, vienen acompañados de rótulos.


Figura 6. Muestras 33 a 40 en la secuencia de montaje ordenada por el contenido de las pistas.
Fuente: Elaboración propia

Parece no haber problema a la hora de describir las muestras 33 y 34. El total que contiene la muestra 33 corresponde con una nueva intervención a cámara de Ricardo Ten, de modo que en principio su voz ya será reconocida por el receptor con problemas de visión y, además, en dicha intervención cuenta cómo le gusta prepararse para la competición, por lo que la incorporación del rótulo a la audiodescripción podría resultar redundante, ya que la identificación del personaje quedaría clara. Incluso parece que se podría añadir brevemente información sobre la entrada de Ricardo en la piscina de la competición.

Por su parte, el total de la muestra 34, que se corresponde con la primera ocasión en la que habla a cámara David Román, el entrenador de Ricardo Ten, se puede adelantar, dado que, tras el total de la muestra 33, se añade un plano recurso (imagen de David ayudando a Ricardo a acabar de colocarse el gorro) que permite disponer de un hueco de mensaje en el que se puede avanzar quién es el siguiente entrevistado.

Las muestras 35 a 40 encadenan seis totales sin ningún plano recurso entre ellos, tres de los cuales se corresponden con personajes nuevos, en el sentido de que no aparecen con anterioridad y se les presenta mediante rótulo. Al enlazar seis intervenciones de personajes a cámara, de cuyas intervenciones únicamente podemos deducir que son deportistas y, en algunos casos, compañeros de Ricardo Ten, resulta casi imposible añadir ninguna unidad descriptiva. Solo entre las muestras 38 y 39 parece haber un hueco de mensaje de un segundo en el que se podría identificar al entrevistado anterior o al siguiente, pero no a ambos.

Se suele pensar que las personas ciegas son buenas reconociendo voces. Sin embargo, la realidad es que tienen más dificultades para recordar qué voz corresponde a qué personaje en obras audiovisuales con multitud de personajes (Fryer 2016), como sería el caso de este documental, en el que intervienen más de 50 entrevistados.

Por tanto, a modo de conclusión preliminar, se podría recomendar evitar el encadenamiento de totales, que supone un problema a la hora de identificar a los personajes. Se podrían añadir planos recurso que permitan incorporar la audiodescripción cuando la información solo esté disponible a través del canal visual o recomendar a los entrevistados que se presenten o den detalles que permitan saber quiénes son, es decir, que se auto audiodescriban, de modo que la audiodescripción del rótulo fuese redundante.

4.3. La combinación de planos recurso y totales con rótulos

Otra de las técnicas narrativas que se suele encontrar con más frecuencia en piezas informativas y de carácter documental es el empleo de planos recurso. Estos planos, que no están directamente relacionados con la acción principal que ocurre en pantalla, permiten completar o ambientar dicha acción.

Las muestras 4 a 8 de nuestro análisis (TCR 00.00.42 a 00.01.21) presentan tres totales con rótulos que se intercalan con planos recurso. Concretamente, los totales con rótulos presentan intervenciones a cámara de la madre, los hermanos y el padre de Ricardo Ten (muestras 4, 6 y 8), entre las que aparecen breves planos recurso de las piernas de Ricardo Ten subiendo al trampolín de una piscina (muestras 5 y 7). Puesto que en este momento del documental aún no se conoce a qué se dedica Ricardo profesionalmente, estos planos recurso funcionan narrativamente a modo de avance o introducción de lo que posteriormente se mostrará con detalle.


Figura 7. Muestras 4 a 8 en la secuencia de montaje ordenada por el contenido de las pistas.
Fuente: Elaboración propia

Durante estos planos recurso encontramos huecos de mensaje breves (de dos y tres segundos) entre total y total. El empleo de planos recurso entre totales puede ser, como hemos visto, muy útil para poder audiodescribir los totales justo antes o justo después de que aparezcan. En el caso que nos ocupa, resultaría muy útil emplearlos para presentar a los familiares de Ricardo Ten, que hablan a cámara por primera vez. El problema radica en que, si tenemos en cuenta la función narrativa de estos planos recurso (aquí, introducir la profesión de Ricardo Ten), también podríamos considerar interesante describirlos, pero el breve hueco de mensaje existente mientras aparecen en pantalla no permite incluir la descripción de los totales y la de los planos recurso. El audiodescriptor se verá obligado a priorizar la información que más haga avanzar la acción y, probablemente, elegirá describir los totales con rótulos en los huecos de mensaje que presentan los planos recurso, perdiendo así la información aportada por estos últimos.


Fotogramas 8, 9, 10, 11 y 12. Secuencia de tres totales con rótulos que se intercalan con planos recurso (TCR 00.00.42 a 00.01.21, muestras 4 a 8)

Tras el análisis se detecta la necesidad de incluir planos recurso de duración suficiente entre total y total para audiodescribir tanto el recurso como el total. La descripción del total podría omitirse si, como proponíamos en el punto anterior, los entrevistados se presentan y dan detalles que permitan saber quiénes son, es decir, se auto audiodescriben.

Si esto no es posible, tendremos que aceptar que, muy probablemente, la descripción de los planos recurso se perderá en favor de la descripción de totales, atendiendo a máximas de relevancia informativa, como sugiere la norma UNE en la definición de ‘bocadillo de información o unidad descriptiva’ (AENOR 2005: 5).

4.4. El encabalgado de audio sobre imágenes

En las muestras 30 a 32, correspondientes al TCR comprendido entre 00.17.43 y 00.18.54, se emplea otra de las técnicas propias del género documental: el encabalgado de audio sobre imágenes.


Figura 8. Muestras 30 a 32 en la secuencia de montaje ordenada por el contenido de las pistas.
Fuente: Elaboración propia

Ricardo Ten acaba de llegar a Glasgow para participar en el campeonato del mundo de natación paralímpica. Mientras en pantalla se muestran planos recurso de Ricardo saliendo del ascensor de su hotel y dirigiéndose hacia la calle para coger un taxi, su voz en off, procedente de una entrevista tapada, cuenta lo ilusionado que está por poder participar en el campeonato (muestra 30). Gracias a que el encabalgamiento del audio se produce un segundo después de que Ricardo aparezca saliendo del ascensor, disponemos aquí de un hueco de mensaje que permitiría una breve audiodescripción de la escena. Posteriormente (muestra 31, TCR 00.18.02-00.18.22), en el taxi, Ricardo observa la ciudad en silencio durante unos segundos, justo antes de que su voz en off, continuación de la anterior entrevista tapada, se vuelva a encabalgar en el 00.18.11 sobre la imagen de los exteriores de Glasgow. Por fin, en 00.18.13, la voz de Ricardo en entrevista tapada y la fuente que lo produce en imagen se unen, y Ricardo continúa ahora hablando a cámara dentro del taxi.

Los segundos de silencio con planos recurso dentro del taxi antes del nuevo encabalgamiento de audio permitirían describir que Ricardo observa la ciudad y que después habla a cámara dentro del taxi. En la siguiente escena (muestra 32, TCR 00.18.23-00.18.54) Ricardo Ten entra al salón del hotel, se sienta en un sofá y llama por teléfono, todo ello con la voz en off de Ricardo procedente de la entrevista que se le estaba realizando antes en el taxi. A partir de 00.18.19, el nadador comienza a hablar por teléfono con su mujer, sin hueco de mensaje entre la voz en off anterior y esta conversación. La inexistencia de hueco dificulta su audiodescripción, aunque el encabalgamiento de ideas (primero, en la entrevista, Ricardo explica que cuando está fuera echa de menos a su familia y los suele llamar por teléfono y, luego, una vez en el sofá del hotel, habla por teléfono y de la conversación se deduce que está hablando con su mujer) ayuda a que parte de la escena sea accesible, es decir, que se auto audiodescriba. Por tanto, la descripción necesaria para entender la escena se reduciría a la mínima expresión y tan solo habría que mencionar, quizá brevemente en el 00.18.28 y aunque se pisara la parte final de la intervención en off, que Ricardo se sienta en un sofá y llama por teléfono.

Tras el análisis se propondría que en estos casos de encabalgamiento de audio o audio split sobre imagen, el sonido inicial se distanciara siempre de la fuente de imagen que lo produce el tiempo suficiente que permita la audiodescripción de la escena. Por otra parte, hemos visto que la inclusión de planos recurso entre totales, por una parte, y que la propia voz de las intervenciones de los personajes auto describa parte de la escena son técnicas que pueden facilitar mucho la labor audiodescriptora.

4.5. El montaje paralelo de imágenes

Entre las muestras 78 y 98 se detecta un montaje paralelo del equipo de gimnasia rítmica que combina el mismo ejercicio de cintas en entrenamientos y en la competición. En esta producción sirve para contar que la excelencia en competición solamente es posible a través del esfuerzo en el entrenamiento y también le da mayor dramatismo porque el resultado de muchos meses de entrenamiento se ve reflejado en unos pocos minutos de competición.


Figura 9. Muestras 78 a 98 en la secuencia de montaje ordenada por el contenido de las pistas.
Fuente: Elaboración propia

En general algunos estándares de audiodescripción recomiendan evitar la terminología específica sobre técnicas cinematográficas (Audio Description Coalition 2009). Si se sigue esta convención, la audiodescripción se centraría en el ejercicio de gimnasia sin detallar el montaje paralelo. Sin embargo, a pesar de que los espectadores parecen tener opiniones contrarias (Fryer 2016), en los últimos años se ha propuesto ir introduciendo descripciones de las técnicas cinematográficas en la AD (Kruger 2010; Fryer 2010; Fryer y Freeman 2013; Perego 2014). Cuando dichas técnicas reflejen el estilo de la realización o tengan un significado metafórico, se pueden describir en la audiointroducción o en la propia audiodescripción con el objetivo de incrementar la alfabetización cinematográfica de los receptores con problemas de visión.

Este fragmento, que combina el montaje paralelo con planos recurso de los familiares en las gradas, se caracteriza por disponer de huecos de mensaje. Apenas hay diálogos (solo los gritos de ánimo de los familiares desde las gradas) y el elemento que cobra importancia es la música, la misma al ritmo de la cual realizan las gimnastas su ejercicio.

Dado el carácter simbólico de este montaje paralelo y debido a la existencia de huecos de mensaje, sería interesante describir la técnica narrativa al principio, con el propósito de centrarse más tarde en el ejercicio en sí que realizan las gimnastas y sin olvidar dejar fragmentos sin audiodescribir para oír los gritos de ánimo de las gradas y la música.

En este caso, parece que no hay problema a la hora de audiodescribir el montaje y quedaría clara la técnica cinematográfica. No obstante, a modo de conclusión preliminar, sería adecuado que en el proceso de montaje se tuviera en cuenta su dificultad y su significado con el propósito de dejar más o menos huecos de mensaje para su descripción.

4.6. El flashback

En las muestras 117 a 122, que coinciden en concreto con los últimos minutos del documental, se reproduce una intervención de Elena Tejedor, directora de la Fundación Trinidad Alfonso. Mediante la técnica de la entrevista tapada, se escucha a Elena Tejedor mientras se reproducen imágenes de los cinco deportistas protagonistas del documental. Estas imágenes coinciden con escenas que ya se han emitido previamente, es decir, se trata de un flashback de imagen (y, en la muestra 122, también de audio), que permite recuperar personajes. A modo de resumen, tiene una gran efectividad en este documental porque permite recordar los valores que simboliza cada uno de los deportistas.


Figura 10. Muestras 117 a 122 en la secuencia de montaje ordenada por el contenido de las pistas.
Fuente: Elaboración propia

La dificultad a la hora de audiodescribir estas imágenes radica en que, al escucharse en todo momento la intervención de Elena Tejedor, no es fácil encontrar huecos de mensaje. Además, al no hacer referencia explícita a las imágenes, lo único que se puede deducir es que la entrevistada representa al proyecto que da apoyo a estos deportistas.

En las ocasiones en las que no hay hueco de mensaje para describir la acción de forma simultánea, los estándares proponen avanzar o posponer la descripción. Tal y como está montado el fragmento, se halla un hueco de mensaje a partir de la muestra 121, cuando termina la intervención de Elena Tejedor. Sin embargo, esto supondría que los receptores ciegos no sabrían quién está hablando hasta que no termine.

Otra posibilidad sería adelantar la descripción de la intervención de Elena Tejedor especificando que se escucha sobre imágenes de los cinco deportistas. En este documental, no es posible, porque no hay hueco de mensaje entre el inicio de su intervención (muestra 117) y la intervención anterior (muestra 116).

De nuevo, como en apartados anteriores, sería pertinente considerar esta cuestión en el proceso de montaje audiovisual y quizás entonces añadir un plano recurso para que diera tiempo a describir lo que va a suceder a continuación.

5. Conclusiones

El objetivo principal de esta investigación era analizar cómo integrar la audiodescripción en el montaje audiovisual para que un producto audiovisual pueda ser apto para todos los espectadores con problemas de visión desde el inicio. En este sentido, de este estudio, centrado en la postproducción del documental Vidas Olímpicas, emerge, en primer lugar, la secuencia de montaje como una herramienta valiosa para el análisis del texto audiovisual, más incluso que el propio producto audiovisual terminado, porque el timeline funcionaría a modo de partitura que permitiría observar los diferentes elementos significativos por separado. Así se detectan intuitivamente y de forma visual aquellos fragmentos más fácilmente audiodescriptibles e incluso aquellos que impedirían una correcta audiodescripción por la inexistencia de huecos. La utilización de la secuencia de montaje puede permitir al profesional del audiovisual poder prever esta circunstancia durante el proceso de montaje.

A modo de conclusión preliminar, queda también patente que el encadenamiento de totales dificulta la audiodescripción, a no ser que los propios entrevistados proporcionen la información que se necesita para identificarlos. En el marco del proyecto de investigación ITACA, se tendrá que comprobar con el estudio de recepción si la incorporación de planos recurso para disponer de hueco de mensaje no repercute en la narración visual. Estos planos recurso quedarían sin audiodescribir, probablemente por cuestiones de relevancia informativa, a menos que se tome la decisión de ampliar su duración.

Asimismo, tras el análisis realizado también hemos detectado otros aspectos del proceso de edición que habría que tener en cuenta para facilitar la accesibilidad de la obra son:

  • Dejar un hueco de mensaje suficiente para audiodescribir el texto en pantalla (se puede alargar la duración de planos que contengan información textual en pantalla, como propone Romero Fresco 2013).
  • En casos de encabalgamiento de audio, distanciar el sonido inicial de la fuente de imagen que lo produce el tiempo suficiente para describir la escena.
  • En secuencias donde se utiliza el flashback de imagen, añadir planos recurso con anterioridad para disponer de hueco de mensaje suficiente para describir la secuencia posterior.

Otro resultado fundamental del presente trabajo es la adopción de una terminología común que sea a su vez aceptada tanto desde el ámbito de la traducción como del audiovisual. Uno de los primeros escollos que encuentra una investigación de carácter interdisciplinar es la diferente nomenclatura utilizada e incluso el distinto conocimiento de la materia que tienen los investigadores. Se da la circunstancia de que cuestiones que para los investigadores del ámbito audiovisual resultan obvias, para los investigadores de la traducción no lo son y viceversa. Resulta muy gratificante comprobar cómo después de varios meses de trabajo conjunto el resultado final puede ser accesible y útil para profesionales y académicos del audiovisual y de la traducción.

Este trabajo compartido también derivó en un modelo de análisis propio que hiciera compatibles los intereses de ambas disciplinas. La colaboración con investigadores del audiovisual facilitó la identificación de los materiales de producción clave en la preparación de un proyecto documental y además facilitó que la productora Anlo participara en la investigación cediendo el material de preparación del proyecto del documental (guion, proyecto, secuencias de montaje etc.)

Las propuestas que derivan de las conclusiones preliminares de este artículo se pondrán en práctica en la elaboración de material audiovisual prevista en el proyecto ITACA (Inclusión social, Traducción audiovisual y Comunicación audiovisual). Como primera aproximación, el director César Martí ya ha implementado algunas de estas consideraciones en su nuevo documental Km 0 (2016).


AENOR (Asociación Española de Normalización y Certificación) (2005) Norma UNE 153020: Audiodescripción para personas con discapacidad visual. Requisitos para la audiodescripción y elaboración de audioguías, Madrid, AENOR.

---- (2012) Norma UNE 153010: Subtitulado para personas sordas y personas con discapacidad auditiva, Madrid, AENOR.

Arnáiz Urquiza, Verónica (2012) Subtitling for the Deaf and the Hard-of-Hearing: Some parameters and their evaluation, tesis doctoral, Universitat Autònoma de Barcelona, Barcelona.

Audio Description Coalition (2009) Standards for Audio Description and Code of Professional Conduct for Describers.

Cabeza Cáceres, Cristóbal (2013) L'audiodescripció de pel·lícules en català: normes, pràctiques, recepció i propostes de millora, tesis doctoral, Universitat Autònoma de Barcelona, Barcelona.

Cerezo Merchán, Beatriz, e Irene de Higes Andino (2013) “Trabajo colaborativo y desempeño profesional: un caso práctico en la clase de accesibilidad audiovisual”, Hikma nº 12: 65–85.

CERMI (Comité Español de Representantes de Personas con Discapacidad) (2015) Informe de seguimiento del subtitulado y la audiodescripción en la TDT. Año 2014, Madrid, Real Patronato de la Discapacidad.

Chaume, Frederic (2004) Cine y traducción, Madrid, Cátedra.

CNMC (Comisión Nacional de los Mercados y la Competencia) (2016) Informe sobre el grado de cumplimiento de las obligaciones impuestas en materia de accesibilidad y presencia de las personas con discapacidad en los medios audiovisuales (2014-2015), Madrid y Barcelona, CNMC.

Cole, Alistair (2015) Good Morning, Grade One. Language ideologies and multilingualism within primary education in rural Zambia, tesis doctoral, University of Edinburgh, Edimburgo.

Crow, Liz (2005) “Making Film Accessible”, URL: http://www.roaring-girl.com/work/making-film-accessible/ (acceso 29 de mayo de 2017)

Directiva 2010/13/UE del Parlamento Europeo y del Consejo, de 10 de marzo de 2010, sobre la coordinación de determinadas disposiciones legales, reglamentarias y administrativas de los Estados miembros relativas a la prestación de servicios de comunicación audiovisual (Directiva de servicios de comunicación audiovisual). Diario Oficial de la Unión Europea, 10 de marzo de 2010.

European Accessibility Act (Propuesta, de 2 de diciembre de 2015, de DIRECTIVA DEL PARLAMENTO EUROPEO Y DEL CONSEJO relativa a la aproximación de las disposiciones legales, reglamentarias y administrativas de los Estados miembros por lo que se refiere a los requisitos de accesibilidad de los productos y los servicios), URL: http://eur-lex.europa.eu/legal-content/ES/TXT/HTML/?uri=CELEX:52015PC0615&from=EN (acceso 29 de mayo de 2017).

Fernández Díez, Federico, y Juan Martínez Abadía (1999) Manual básico de lenguaje y narrativa audiovisual, Barcelona, Paidós.

Fernández Fernández, Cesáreo, y Esteban Galán (2011) “La televisión pública estatal en España: entre el apagón analógico y el apagón publicitario” en La televisión pública a examen, Bienvenido León (coord.), Madrid, Comunicación Social Ediciones y Publicaciones: 162–173.

Fox, Wendy (2016) “Integrated Titles - An Improved Viewing Experience? A Contrastive Eye Tracking Study on Traditional Subtitles and Integrated Titles for Pablo Romero-Fresco’s Joining the Dots”, en Eyetracking and Applied Linguistics, Silvia Hansen-Schirra y Sambor Grucza (eds.): 5-29.

Fresno Cañada, Nazaret (2014) La (re)construcción de los personajes fílmicos en la audiodescripción. Efectos de la cantidad de información y de su segmentación en el recuerdo de los receptores, tesis doctoral, Universitat Autònoma de Barcelona, Barcelona.

Fryer, Louise (2010) “Directing in reverse” en The changing face of translation. Proceedings of the 9th Annual Portsmouth Translation conference, Ian Kemble (ed.): 62–72.

---- (2016) An Introduction to Audio Description. A Practical Guide, Londres y Nueva York, Routledge.

----, y Jonathan Freeman (2013) “Cinematic language and the description of film: keeping AD users in the frame”, Perspectives 21, nº 3: 412–426.

Galán, Esteban (2008) Televisión en virtual, Madrid, IORTV.

Gómez López, Antía (2013) “La escritura del deseo en la publicidad televisiva”, Escritura e Imagen, nº 9: 51-82.

Kruger, Jan-Louis (2010) “Audio-narration: re-narrativising film”, Perspectives 18, nº 3: 231–249.

Lambourne, Andrew (2012) “Climbing the Production Chain”, presentación en la conferencia 8th Languages and The Media, Berlín, 21–23 de noviembre de 2012.

Ley 7/2010, de 31 de marzo, General de la Comunicación Audiovisual. Boletín Oficial del Estado, 1 de abril de 2010.

Mace, Ronald (1988) Universal Design: Housing for the Lifespan of All People, Washington, D4C., U4S4 Department of Housing and Urban Development.

Marzal-Felici, Javier, y Francisco Javier Gómez-Tarín (2015) Diccionario de términos y conceptos audiovisuales, Barcelona, Cátedra.

----, y Longi Gil (eds) (2008) Eines per a la producció de vídeo documental, Benicarló, Onada Edicions.

Maszerowska, Anna, Anna Matamala y Pilar Orero (eds) (2014) Audio Description. New perspectives illustrated, Ámsterdam, John Benjamins.

Matamala, Anna (2014) “Audio describing text on screen” en Audio description: new perspectives illustrated, Anna Maszerowska, Anna Matamala y Pilar Orero (eds.), Ámsterdam, John Benjamins: 103–120.

----, y Pilar Orero (eds) (2009) Listening to Subtitles. Subtitles for the Deaf and Hard of Hearing, Berna, Peter Lang.

----, y Pilar Orero (2015) “Text on screen” en Pictures painted in words: ADLAB Audio Description guidelines, Aline Remael, Nina Reviers y Gert Vercauteren (eds.), Trieste, Edizioni Università di Trieste: 39-42.

----, y Pilar Orero (eds) (2016) Researching Audio Description, Londres, Palgrave Macmillan.

Muller, Tia (2016) A user-centred study of the norms for subtitling for the deaf and hard-of-hearing on French television, tesis doctoral, Universitat Autònoma de Barcelona, Barcelona.

ONU (Organización de Naciones Unidas) (2006) Convención sobre el derecho de las personas con discapacidad.

Perego, Elisa (2014) “Film Language and Tools” en Audio description: new perspectives illustrated, Anna Maszerowska, Anna Matamala y Pilar Orero (eds.), Ámsterdam, John Benjamins: 81–102.

Porteiro Fresco, Minia (2012) El subtitulado como herramienta complementaria en la rehabilitación logopédica de patologías lingüísticas, tesis doctoral, Universitat Autònoma de Barcelona, Barcelona.

Remael, Aline, Nina Reviers y Gert Vercauteren (eds) Pictures painted in words: ADLAB Audio Description guidelines, Trieste, Edizioni Università di Trieste.

Romero Fresco, Pablo (2013) “Accessible filmmaking: Joining the dots between audiovisual translation, accessibility and filmmaking”, JoSTrans nº 20: 201–223.

---- (ed) (2015) The Reception of Subtitles for the Deaf and Hard of Hearing in Europe. UK, Spain, Italy, Poland, Denmark, France and Germany, Berna, Peter Lang.

Tamayo Masero, Ana (2015) Estudio descriptivo y experimental de la subtitulación en TV para niños sordos. Una propuesta alternativa, tesis doctoral, Universitat Jaume I, Castelló de la Plana.

Udo, John-Patrick, y Deborah Fels (2009) “The Rogue Poster Children of Universal Design: Closed Captioning and Audio Description”, Ted Rogers School of Information Technology Management Publications and Research nº 18: 1–32.


Km 0 (César Martí 2016)

Vidas Olímpicas (César Martí 2015)


[1] En este artículo se hará referencia a la subtitulación para sordos a través de su sigla en español, SPS. Es habitual encontrarlo en inglés bajo las siglas SDH (Subtitling for the D/deaf and Hard-of-hearing).

[2] Queda fuera del alcance de este trabajo la accesibilidad de productos audiovisuales para personas con discapacidad intelectual o con dificultad lectora (cf. Porteiro Fresco 2012).

[3] En Europa guiadas por la Directiva 2010/13/UE, de servicios de comunicación audiovisual, del Parlamento Europeo y del Consejo de 10 de marzo de 2010 sobre la coordinación de determinadas disposiciones legales, reglamentarias y administrativas de los Estados miembros relativas a la prestación de servicios de comunicación audiovisual (en proceso de revisión desde 2016). Disponible en: http://eur-lex.europa.eu/legal-content/ES/TXT/PDF/?uri=CELEX:32010L0013&from=ES.

[6] En el año 2010 las cadenas españolas (públicas y privadas) subtitularon un promedio del 37 por ciento de sus emisiones televisivas, porcentaje que creció hasta el 67 por ciento alcanzado en el año 2014 (CERMI 2015) y hasta el 73,8 por ciento según datos del año 2015 (CNMC 2016). En el caso de la audiodescripción, las cadenas españolas (públicas y privadas) audiodescribieron unas 27 horas en todo el año 2011, más de 120 horas en 2014 (CERMI 2015) y alcanzaron una media mensual de 31 horas en 2015. Como vemos, los porcentajes promedio pretendidos por ley para el año 2013 no han llegado a alcanzarse en ninguna de las dos modalidades. Aun así, podemos decir que la evolución de los niveles de programación accesible ha sido favorable hasta la fecha.

[9] Disponible en: http://www.whatscine.es/.

[10] Además de España, otros países como Reino Unido, Francia, Alemania, etc. también cuentan con normativas nacionales que regulan la práctica de la SPS y la AD. Asimismo, a día de hoy también contamos con referencias a nivel internacional, como la interesante propuesta de convenciones para la AD del proyecto ADLAB (Remael, Reviers y Vercauteren 2015, http://www.adlabproject.eu/Docs/adlab%20book/index.html).

[13] A pesar de esta invisibilidad tradicional de la traducción audiovisual, es destacable el creciente desarrollo de servicios accesibles fuera de la industria cinematográfica clásica, en plataformas como Netflix, Filmin, etc., lo que, de algún modo, supone una ventana de esperanza para la TAV.

[14] En otros países, como es el caso de España, su introducción se encuentra aún en una fase incipiente. De modo exploratorio, en 2013 miembros del Departament de Traducció i Comunicació de la Universitat Jaume I iniciaron una colaboración con el Servei de Comunicació de la Universitat Jaume I para la elaboración de vídeos accesibles para el blog de divulgación CienciaTV de esta universidad (http://blogs.uji.es/cienciatv). También trabajan desde el enfoque del accessible filmmaking autores como, entre otros, Crow (2005), Cole (2015) y Fox (2016).

[15] Esto se encuentra en línea con lo afirmado por el Informe de seguimiento del subtitulado y la audiodescripción en la TDT 2014 (CERMI 2015: 127), según el cual ‘los niveles de audiodescripción requeridos por la ley [en España] son ampliamente inferiores a los impuestos al subtitulado [para personas sordas], en parte debido a la mayor dificultad para la audiodescripción de determinados tipos de programa’.

[17] El pitching es una técnica que se emplea en la producción audiovisual al objeto de captar la atención de productores, socios o colaboradores mediante la exposición de los aspectos más relevantes de la obra audiovisual. El anglicismo pitch proviene del mundo del béisbol y se emplea como metáfora en la que el profesional que defiende el proyecto sería el pitcher que, como en el béisbol, basa su destreza en lanzar e impactar en los receptores con potencia y claridad.

[18] Seleccionar el inicio de una determinada línea narrativa permite acceder a las elecciones estilísticas que después se repiten en el resto de personajes. Tal vez por el hecho de ser el primer personaje, el tratamiento visual que se realiza es siempre más canónico, es decir, el que mejor representa el estilo visual del documental (Marzal y Gil 2008).

[19] Otra de las razones que motivó también la elección de este documental fue justamente el mensaje de inclusión que transmite la obra audiovisual reivindicando que el deporte iguala a todos; por eso, se da el mismo tratamiento a los deportistas olímpicos y paralímpicos.

About the author(s)

Beatriz Cerezo Merchán Holds a BA and a PhD in Translation and Interpreting. At present she is a lecturer of translation and English language at the Universitat de València. She is a member of the research groups CiTrans (http://citrans.uv.es/) and TRAMA (http://www.trama.uji.es/), and her main areas of research are the didactics of translation, and audiovisual translation and accessibility.

Irene de Higes Andino (BA and PhD in Translation and Interpreting) is lecturer of the Translation and Communication department at Universitat Jaume I and member of the research group TRAMA (http://www.trama.uji.es). She has worked as a production assistant in a dubbing studio and as a freelance translator specialised in audiovisual translation. Her research interests focus on multilingualism, identity, audiovisual translation and accessibility.

Esteban Galán is a researcher of the group “Technologies Applied to Audiovisual Communication” (ITACA) in the Jaume I University of Castellón, Spain. He has worked as a post-production technician and he has more than 15 years of experience as audiovisual producer in different TV broadcasters. He has been awarded with several academic prizes as the 3rd National Award (2003) and he has articles, courses, lectures and international conferences about Transmedia, TV Technologies and Media. He is fluent and has international teaching experience in English and French. Since 2017 he runs the podcast Transmedia: el programa de investigación en comunicación.

Roberto Arnau Roselló has a degree in Audiovisual Communication from the University of Valencia (1997) and a PhD from the Jaume I University of Castellón (2006). He has taught at the University Institute of Technology of the University of Franche Comté (Besançon, France) where he has carried out technological projects, teaching and research tasks. He is currently a professor at the Universitat Jaume I in the degrees of Audiovisual Communication and Journalism, as well as in the University Master in New Trends and Processes of Innovation in Communication. Director of the Laboratory of Communication Sciences (LABCOM) from 2006 to 2016, his activity has been developed both in the field of research, as in management, and in the professional practice in the audiovisual sector (as a scriptwriter, director , producer, director of photography or assistant director). Researcher in the research group ITACA-UJI (Research in technologies applied to audiovisual communication), among his publications include book chapters and communications in international conferences, as well as articles in scientific journals such as Studies on the Journalistic Message, Fotocinema, Historia y Social Communication, ICONO 14, or Communication and Media.

Email: [please login or register to view author's email address]

©inTRAlinea & B. Cerezo Merchán, I. de Higes Andino, E. Galán, & R. Arnau Roselló (2017).
"Montaje audiovisual e integración de la audiodescripción en la producción documental"
inTRAlinea Special Issue: Building Bridges between Film Studies and Translation Studies
Edited by: Juan José Martínez Sierra & Beatriz Cerezo Merchán
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2252

Accessible Filmmaking in Documentaries

By Pablo Romero Fresco (Universidade de Vigo, Spain and University of Roehampton, UK)

Abstract & Keywords

In line with current attempts to build bridges between film and translation, the present article applies the notion of accessible filmmaking (the integration of audiovisual translation and accessibility as part of the filmmaking process) to the theory and practice of documentary film. The first part of the article offers a brief historical overview of this genre, with particular emphasis on its key defining features and a discussion of how the translation of documentaries has been handled so far by filmmakers and translators. After a short introduction about the progress made up to date regarding the training, research and practice of accessible filmmaking, the second and main section of this article explores how this approach has been implemented in four recent documentaries: Joining the Dots (2012), A Grain of Sand (2014), Colours of the Alphabet (2016) and Notes of Blindness (2016). These films are used here to illustrate how translation and accessibility can be integrated in the film production workflow, to discuss the political and ethical impact of the use of subtitles in ethnographic documentaries and to analyse the role of translation and accessibility in the (pre)production and post-production stage, with special emphasis on the use of integrated titles. The early consideration of translation and accessibility in documentary films is presented here as an issue of ethics and responsibility for filmmakers vis-à-vis the participants in the film and its foreign and sensory-impaired viewers.

Keywords: accessible filmmaking, documentary, integrated titles, audiovisual translation

©inTRAlinea & Pablo Romero Fresco (2017).
"Accessible Filmmaking in Documentaries"
inTRAlinea Special Issue: Building Bridges between Film Studies and Translation Studies
Edited by: Juan José Martínez Sierra & Beatriz Cerezo Merchán
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2251

1. Introduction

Quality is nowadays a key concern for researchers, trainers and professionals working in the field of audiovisual translation (AVT) and media accessibility. Yet, many of the efforts made to assess and increase quality in these areas are often thwarted by the currently prevailing model of industrial AVT (Pérez-González 2012). In this model, translation and accessibility are relegated to the distribution stage of the filmmaking process and outsourced to translators who work with tight deadlines, small remuneration and no access to the creative team of the films. This divorce between film(making) and translation is also a reflection (or perhaps the main cause) of the gap between film studies and AVT studies, two disciplines that have so far interacted much less than what could be expected given their overlaps and shared interests. Over the past years, different attempts have been made to bridge this gap and explore the potential for cross-fertilisation between the two areas (Chaume 2004, Díaz-Cintas 2004, Remael 2004, Nornes 2007, Cattrysse and Gambier 2008, Martínez Sierra 2012, Bosseaux 2015, Redmond and Batty 2015). This volume is further proof of this promising – if still incipient – trend, which can contribute to put forward a wider and more inclusive view on film. In an increasingly multilingual world, a monolingual approach to film studies is bound to leave behind the largest part of the viewers, not only foreign and sensory-impaired audiences but also viewers of the growing number of films that include more than one language in their original versions.

Against this background, the notion of accessible filmmaking (Romero-Fresco 2013), which proposes the integration of AVT and accessibility as part of the filmmaking process through collaboration between filmmakers and translators, offers an alternative approach to the current model of industrial AVT. The present article discusses this approach as it applies to the theory and practice of documentary filmmaking, a genre that offers great potential for collaboration between film and translation. The first two sections include a brief historical overview of documentary as a genre, its key defining features and how its translation has been approached so far from the point of view of filmmakers and translators. This is followed by a brief account of the progress made so far in the training, research and practice of accessible filmmaking. Finally, the main section of the paper presents a discussion of how translation and accessibility, with a special focus on subtitling, have been integrated into the (pre-)production and post-production stages of four different documentaries that were made following the principles of accessible filmmaking.

2. Documentary filmmaking: a brief overview

The first documentary films were made long before the term ‘documentary’ was coined to refer to this genre. Some of the earliest examples that started to define the key elements of this type of film were the Lumière brothers’ Train Pulling into a Station (1895) and Workers Leaving The Lumière Factory in Lyon (1895), which allegedly prompted a member of the audience to shout ‘It’s life itself!’ (De Jong, Knudsen and Rothwell 2011: 340). In 1919, Russian filmmaker Dziga Vertov published his manifesto Kinoks-Revolution Manifesto, calling for a cinema that documents real life and reports truth. Three years later, Robert Flaherty made Nanook of the North (1922), widely regarded as the first feature-length documentary and one of the first ethnographic films ever made. Flaherty’s film includes some of the key characteristics of this genre: third-person narration, a subjective tone and the focus on the indigenous ‘others’. In 1926, John Grierson coined the term ‘documentary’ in his review of Flaherty’s second film, Moana (1926), thus setting the official birth of the genre (Chanan 2007).

Soon, documentary filmmakers began to explore the boundaries between reality and fiction, reflected in the tension between Dziga Vertov’s realistic credo ‘life caught unawares’ and the more intrusive, life-arranged-for-the-camera approach adopted in his classic Man with the Movie Camera (1929) (Taylor 1998: 75). Further examples of the latter approach are the propaganda documentaries made by the US government in the 1930s, the tribute to the Nazi Party made by the acclaimed yet repudiated Leni Riefenshtal in Triumph of the Will (1934) and Luis Buñuel’s Land without Bread (1933), which combined surrealism and political activism to blur the lines between documentary and fiction. In the 1950s and 1960s, Vertov’s realistic ‘life caught unawares’  was put into practice by the emergence of direct cinema (US), cinéma verité (France) and free cinema (Canada and UK). The appearance of new lightweight cameras and synchronised sound allowed filmmakers to develop a new observational style that rejected rehearsals, interviews, commentaries, film lights and in general any sort of intrusion (De Jong, Knudsen and Rothwell 2011). Jean Rouch applied these principles to ethnographic documentaries in order to ‘show people without masks, without make-up, to catch them through the eye of the camera in a moment when they are not acting, to read their thoughts, laid bare by the camera’ (Vertov 1984: 41). In the 1960s and 1970s, documentary filmmakers joined the new civil rights, anti-war, women’s and gay rights movements, replacing the detached observational nature of cinéma vérité with a more personal and engaged approach. The introduction of camcorders favoured the production of more subjective documentaries and the spread of television provided a new home for this genre (De Jong, Knudsen and Rothwell 2011).

Since then, documentaries have embraced a wide variety of themes and production approaches, and they have been subdivided in different ways, including Nichols’ classic distinction between expository, poetic, observational, participatory, reflexive and performative modes (Nichols 2001). More recently, the turn of the century has brought the so-called New Documentary wave (Chanan 2007), characterised both by the return of the documentary to the big screen and, at the same time, the spread of affordable digital technology and digital platforms that make it easier to produce and disseminate independent documentaries:

While television spends its budgets on house makeover and reality shows, documentary makers need to look elsewhere for funds. In doing so they get greater autonomy and artistic freedom compared with their television colleagues. (…) Real life doesn’t need a script, just good editing. (The Guardian 2004)

The brief overview of some of the key developments in documentary filmmaking outlined in this section may help to explain why this genre is particularly suited to embrace an accessible filmmaking approach: the wide range of themes covered, the independence and creativity with which it can be produced, often free from the constraints of the big-budget movie industry, the commitment to show the life, culture and language of ‘the other’ as it is, etc. The four documentaries analysed in this article illustrate this variety of approaches, from the life-caught-unawares, ethnographic approach of The Colours of the Alphabet (2016) to the more innovative, docufiction style of Notes on Blindness (2016), they all conform to the age-old definition of documentary as the ‘creative treatment of actuality’ (Grierson 1933: 8). Indeed, real life may only need good editing, and not a script, to become a documentary, but it also needs translation. The next section explores how this has been tackled so far both by filmmakers and by the translation industry.

3. The translation of documentaries

The translation of documentaries varies greatly depending on whether it is handled by filmmakers or outsourced to professional translators. The first approach has been occasionally covered by film studies, whereas the literature on AVT has focused almost exclusively on the second approach, thus illustrating the above-mentioned gap between the two disciplines.

3.1. Translation by filmmakers

Proof of the scarce attention devoted to translation in the literature on documentary film(making) may be found in the absence of any reference to translation in seminal books about the theory and practice of documentary such as Producing with Passion (Fadiman and Levelle 2008) or Creative Documentary (De Jong, Knudsen and Rothwell 2011). An interesting case in point is Michael Rabiger’s Directing the Documentary (2004), which aims to equip the readers with the necessary tools to produce and distribute their documentaries. Arguably, translation is one more of the tasks that these aspiring low-budget filmmakers will have to deal with, directly or indirectly. Yet, of its very comprehensive 648 pages, only half a page is devoted to the production of subtitles, offering the following tips: find a literate native speaker to do the translation, pick a clear typeface in yellow with black edging, compress dialogue to essentials, do not subtitle over shots and use ellipsis between subtitles (ibid: 252).

An exception to this rule is that of ethnographic documentary filmmakers, whose work, largely overlooked in AVT, provides a key contribution to our understanding of the relationship between film and translation and the first example of accessible filmmaking from the point of view of research and practice. As pointed out by Barbash and Taylor (1997: 420), ethnographic documentaries involve a triple case of translation: rendering aspects of one culture intelligible to another, transforming cultural elements into the film medium and transferring meaning from one language into another. In this type of cross-cultural filmmaking, they argue, filmmakers are brokers of meaning and translation is a cultural act involving interpretation, negotiation and, inevitably, compromise:

As cross-cultural film makers, we have to ask ourselves: How much do we feel compelled to translate, allowing that some meaning will always be lost? How can we fill in the gaps of understanding? And how, if at all, can we reveal to the spectators our role in translation and interpretation for what it is? (ibid. 421)

Up until the early 1960s, ethnographic documentaries such as those made by Jean Rouch had normally resorted to dubbing and especially voice over for translation. Filmmakers valued that the viewing experience remained ‘one of looking and listening’ (ibid: 425) and that this mode allowed the translation of several voices at the same time. However, they found this translation mode problematic in transferring the authenticity, rhythm, inflections, nuances and background information of the participants in the film. The pioneering use of subtitles by filmmakers John Marshall and Tim Asch in their documentaries at the beginning of that decade is regarded as a landmark in documentary filmmaking (Henley 2010). It helped filmmakers to construct a closer relationship with their participants and it is also key for AVT studies, since it triggered some of the first theoretical reflections on the political and ethical reasons for the use of subtitles:

What we wanted to replace was not a narrative view of life, but the word-dominated structures of the illustrated lecture film and the all-knowing eye of Hollywood. This resulted in part from our having watched foreign feature films. The people in these films spoke other languages and came from other cultures, but they were still portrayed as individuals. There was no voice on the soundtrack telling you what to think about them. We read their conversations in subtitles and, guided by the filmmaker, we made an analysis of their motivations and actions. (MacDougall 2001-2002: 88)

Unlike in fiction films, subtitles in ethnographic films were regarded from the beginning as one of the creative ingredients of the filmmaking process, a ‘dramatic component of visual anthropology’ (Ruoff 1994) that required to be tackled collaboratively, often by filmmakers and non-professional translators (Lewis 2004), and from the editing stage (Henley 1996). All these elements, as well as the consideration by these filmmakers and scholars of the effect of subtitles upon the audience account for the pioneering role played by ethnographic film studies in the research and practice of accessible filmmaking. 

Subtitling is considered here as a linguistic, cultural and technical challenge, but one that encourages the creativity of the filmmakers and that allows them to develop new meanings and interpretations for their films (Zhang 2012). For MacDougall (1998), the main disadvantage in the use of subtitles is that the dialogue is packaged and acquires a somewhat prophetic nature, viewers lose freedom and become word-focused and filmmakers have to go against the ‘show, don’t tell’ formula that is traditionally recommended to enhance the visual nature of film. Yet, he points out that subtitles also present a great deal of benefits to filmmakers, as they can contribute to further the characterisation of the participants in the film and can help filmmakers to recontextualise, focus or narrow down their ideas. Subtitles are a ‘stamp of possession’ (MacDougall 1998: 174) on a film, the ‘textual eyes’ (Zhang 2012: 447) that allow filmmakers to project their particular interpretation and to speak to the audience while the participants in the film are speaking to each other.

More specifically, many of these scholars reflect on the impact subtitling has on the filmmaking process, anticipating some of the issues that will be discussed in this article. As far as production is concerned and from a practical point of view, subtitling requires key changes in camera framing that must be planned before shooting (Henley 2010), whereas from a more theoretical standpoint subtitles are thought to act on the verbal level as the camera acts on a visual level ‘to single out subjects and frame human relationships’ (MacDougall 1998: 169). As far as editing is concerned, MacDougall also stresses the key role played by subtitles on the rhythm, cadence and tempo of the film. Rhythm in film is thus not only determined by the pace of the editing, an aspect that has been researched exhaustively by film scholars, but also by the pace at which the viewers read the subtitles (Romero-Fresco 2018). This is a consideration that has so far been overlooked in film studies and AVT alike. Thus, the density, speed, exposure time and complexity of the subtitles may be critical to determine the rhythm of a film, which may be different in its translated version. The following quote by MacDougall (1998: 168) sums up the role played by subtitling in ethnographic film:

The writing and placing of subtitles involves considerable polishing and fine-tuning, but unlike the ex post-facto subtitling of a feature film, this remains part of the creative process, influencing the pacing and rhythm of the film as well as its intellectual and emotional content.

Finally, ethnographic filmmakers have also discussed the use and implications of non-subtitling, an issue that is nowadays commonly debated in AVT research in the light of the increasing production of multilingual films that have this option at their disposal (Krämer and Eppler forthcoming). Non-subtitling is used in ethnographic films to enhance the visual nature of the viewing experience and to place the focus on people’s physiognomy, non-verbal forms of interaction and the language of music, which ‘need not be reduced to the mere function of communicating meaning’ (Trinh 1992: 114). Then again, non-subtitling is also acknowledged as a controversial decision, and one that can be seen as self-indulgent, ethnocentric, abusive towards the film participants and orientalist (Barbash and Taylor 1997: 429).

At any rate, it would seem that just as the emergence of ethnographic filmmaking helped to give a voice to silent communities around the world mainly through the use of subtitles, it also contributed to raise the visibility of translation amongst an admittedly small number of filmmakers and film scholars. Unfortunately, the obvious connection between ethnographic filmmaking and AVT studies has until now never materialised, as shown (to mention one example) by the absence of ethnographic filmmakers in AVT publications and conferences such as Media for All and Languages and the Media and by the fact that ethnographic conferences and panels such as the 2011 ‘Subtitling Ethnographic Films: Knowledge and Value in Translation’ have so far fallen off the AVT radar. Given the current relevance of abusive (Nornes 2007), creative (McClarty 2012) and integrated (Fox 2016a) subtitling[1], the use of, and reflection on, subtitling by ethnographic filmmakers is now more relevant than ever and it shows that it is possible to integrate AVT as a creative element within the filmmaking process through the collaboration of filmmakers and translators.

3.2. Translation by professional translators

AVT studies have shown an increasing interest in the translation of documentaries over the past 15 years. This is regarded as a highly versatile discipline that requires ‘an all-round knowledge of text types and functions, language registers, domain-specific terminology and translation techniques such as voiceover, dubbing, off-screen dubbing and subtitling’ (Hanoulle, Hoste and Remael 2014: 359). Dubbing and voice-over are normally the modes chosen for the translation of documentaries in dubbing countries such as Spain, France, Italy and Germany. This has been thoroughly researched by Franco, Matamala and Orero (2010) and more specifically by Franco (2001) with regard to terminological and conceptual issues, and by Orero (2006) regarding synchronisation. Subtitling, the mode used for the translation of documentaries in countries such as Portugal, Belgium or Sweden, has been studied, amongst others, by Ferreira (2002) and Kaufmann (2008). Regardless of the translation mode chosen, this discipline involves a series of challenges, which Matamala (2010: 269) describes as

Identifying terms, understanding terms, finding the right equivalent, dealing with the absence or the inability to find an adequate equivalent, dealing with denominative variation, choosing between in vivo and in vitro terminology and avoiding wrong transcriptions.

Most of this research on the translation of documentaries, whether descriptive or experimental, focuses on translation as an element included in the distribution stage, detached from the creative team of the film. However, as will be argued in this article, some of the challenges identified by Matamala may be tackled successfully from an integrated standpoint that considers translation as part of the filmmaking process. A good example is the last challenge: the use of wrong transcriptions. Matamala (2009: 101) explains the process of translating documentaries as follows:

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.

In order to convert the oral input into a translation script for voice-over or into a written output for subtitles, the translators may or may not have at their disposal a transcript of the original. Often, and especially in the case of same-language subtitles for the deaf and hard of hearing, translators start off by creating this transcript. Here, a wider view of translation as one more element in the overall documentary production reveals a serious case of duplication of efforts. Below is an account provided by Lukasz Daniluk, professional transcriber, of how documentaries are often handled at the post-production stage in the UK:

Documentaries with a great deal of footage often require transcription so that the edit producer can find the best sounding lines in the footage without the need to view it all. This transcription is done manually. The footage is then assembled into a rough cut, modified several times, scrutinised by the executive producers and the channel commissioner and sent to ‘picture lock’, where graphics, colour grading and narration are added. The film is then sent to the broadcaster for transmission and to subtitling companies, where translators often transcribe the documentary from scratch. The original transcripts, created by ‘loggers’ at the beginning of the process, are usually put in recycling bins at the post production facility. (Daniluk 2016, personal communication)

It stands to reason that a more holistic approach to translation could save some of the work currently involved in this process. Firstly, the transcription of the original footage at the beginning of the post-production stage could be done through respeaking instead of manually, which is supported by the positive results obtained by speech-recognition-based transcription in recent studies (Matamala et. al 2017). Secondly, if this transcription was performed by translators (some of whom are already trained as respeakers), they would be able to become familiarised with the documentary at an early stage and, once the final edit is ready for distribution, prepare the template for subtitles and SDH, which can then be used as a basis for translation into other languages. For this to happen, it is necessary to introduce the notion of translation and accessibility in the (post-)production process, as discussed and illustrated in the following two sections.

4. Accessible filmmaking

The current industrial model that relegates AVT and accessibility to the end of the distribution process as add-ons is a profitable one for film producers. Often, the translated and accessible versions of a film make up as much as 60 per cent of its revenue for a cost of between 0.1 per cent and 1 per cent of the total film budget (Lambourne 2012). However, this model can also have a negative impact on quality. Filmmakers such as Ken Loach are now beginning to complain that their films’ vision is sometimes altered in translation and that they have little power to change it or, more worryingly, little knowledge that this is happening (de Higes Andino 2014). In an attempt to propose a different model, accessible filmmaking (Romero-Fresco 2013) aims to integrate AVT and accessibility as part of the filmmaking process, which requires the collaboration between the creative team of the film and the translator. Put it another way, accessible filmmaking involves the consideration during the filmmaking process (and through collaboration between the translator and the creative team of the film) of some of the aspects that are required to make a film accessible to people that cannot, or cannot properly, access it in its original form, including viewers in other languages and viewers with hearing or visual loss. Since it was first introduced, this approach has been applied to training, research and professional practice.

Traditionally, film(making) courses have disregarded translation and accessibility issues, while postgraduate programmes in AVT do not normally include film(making). This is beginning to change, as postgraduate courses in filmmaking such as the Film Studies Masters at the University of Malta and the MA in Film Production at the ESCAC (Barcelona), the leading film school in Spain, now include classes on AVT and accessibility. Likewise, AVT courses are beginning to open the door to film-related contents, as shown by the collaboration between the Universitat de Vic and the ESCAC, the MA in Multimedia Translation at the Universidade de Vigo and especially the MA in Accessibility and Filmmaking at the University of Roehampton, where students learn not only how to make films but also how to make them accessible to viewers in other languages and viewers with hearing and visual loss.

As far as professional practice is concerned, the first examples of accessible filmmaking are almost as old as cinema itself. They can be traced back to silent films, where intertitles were produced as part of the post-production process, often supervised by the filmmakers (Izard 2001), or to multiple-language versions (Vincendeau 1999), where films were made and remade in different languages and translation was an integral part shaping the production process. Since then, translation has only been integrated in the filmmaking process occasionally (such as in the case of ethnographic documentaries), but there are now promising signs showing a changing trend. Recently, and partly due to the emergence of multilingual films, more and more filmmakers are beginning to engage with translation from the production process and to collaborate with translators, as is the case of John Sayles (Lone Star 1996; Men with Guns 1997), Jim Jarmusch (Mystery Train 1989; Night on Earth 1991), Danny Boyle (Slumdog Millionaire 2008), James Cameron (Avatar 2009) and, more notoriously, Quentin Tarantino (Inglourious Basterds 2009) and Alejandro González Iñárritu (Babel 2009; The Revenant 2015), both of whom issued translation guidelines to their distributors in order to ensure that their vision for their films was maintained in the target versions (Sanz 2015)[2]. However, given the inflexible nature of industrial translation, where distributors have the power to decide against the translation wishes of recognised filmmakers such as Ken Loach and Quentin Tarantino (ibid.), independent filmmaking offers an ideal platform for accessible filmmaking to be developed. This is the case of recent independent films that have integrated translation and accessibility from an early stage, such as Michael Chanan’s Secret City (2012), Enrica Colusso’s Home Sweet Home (2012), Elisa Fuksas’ Nina (2012), Alastair Cole’s The Colours of the Alphabet (2016) or the Emmy award-winning Notes on Blindness (Spinney and Middleton 2016).

Interestingly, many of these filmmakers, when faced with the need to use subtitles in the original versions of their films or to engage with translation, have opted for the use of creative or integrated (sub)titles, thus playing with non-standard fonts, display modes, effects or positions in order to fulfil both a linguistic and an aesthetic function in the film (McClarty 2012). Linked to this, another factor that is contributing to the consideration of translational issues early on in the production process is the increasing presence of on-screen text in original films and series. This applies to the creative use of on-screen titling in TV series such as Sherlock (Gatiss and Moffat 2010-) but also to other fiction films, documentaries and series where the nowadays pervasive use of text-based communication is shown by means of on-screen text. It becomes obvious that the use of standard post facto subtitling for these films will create too much of a gap between the original viewers’ experience and that of foreign and sensory-impaired audiences, which can only be solved by working on translation/accessibility at an early stage. Part of the research conducted so far in the field of accessible filmmaking, which will be discussed in the following section, is exploring precisely how subtitling can be integrated as one more element in the mise-en-scene of the film and how this is received by the viewers in terms of opinion, comprehension, visual perception and sense of engagement (Kruger, Doherty and Soto Sanfiel 2015, Fox 2016a).

All in all, the progress made so far in the integration of translation and accessibility as part of the filmmaking process shows that this endeavour is made easier when films are independent (or where the filmmaker has a final say), have a multilingual element and/or resort to abundant (and even creative) use of on-screen text. This explains why documentaries are particularly suited to accessible filmmaking, as will be shown by the four films discussed in the following section. Another little-discussed advantage in the use of independent documentaries (and independent films in general) is that it can encourage AVT studies to move away from the mainstream films that have dominated the material used for research and training in this field since its origins. This mainstream content often presents multiple copyright issues and offers little chance of interaction with its filmmakers, thus perpetuating the gap between film and translation.

5. Accessible filmmaking in documentaries  

The four documentaries discussed in this section are Joining the Dots (2012), A Grain of Sand (2014), Colours of the Alphabet (2016) and Notes on Blindness (2016). Joining the Dots (Romero-Fresco, 2012) is a short participatory documentary that tells the story of Trevor, who lost his sight at the age of 60. Following a period of depression, he found his way out with the help of audiodescription, which also enabled him to rekindle his passion for cinema and theatre. A Grain of Sand (2014) is a short observational documentary made by Kate Dangerfield and Arianna Avruscio in Ciudadela Sucre, Soacha (Colombia), about the construction of a playground in an urban village that suffers the consequences of a social crisis and the result of 50 years of armed conflict. Colours of the Alphabet (2016) is an 80-minute observational documentary directed by Alastair Cole that follows three Zambian children and their families as they enter the formal education system for the very first time in a class where they speak one language, the teacher another and together they must embrace the country’s only official language, English. Notes on Blindness (2016) is a creative docufiction about the theologist John Hull, who, after losing his sight in 1983, began keeping an audio diary that enables the film to explore loss, rebirth, renewal and the interior world of blindness. The film was directed by James Spinney and Pete Middleton and subtitled with integrated titles by Wendy Fox and the author of this article.

Colours of the Alphabet will be used here to illustrate how translation and accessibility can be integrated in the film production workflow and to discuss the political and ethical impact of the use of subtitles in ethnographic documentaries. Joining the Dots, A Grain of Sand and Notes of Blindness will be analysed in relation to how accessible filmmaking can be implemented in specific terms at the (pre)production and post-production stage, with special emphasis on the use of integrated titles.

5.1. A new accessible workflow and its ethical implications

Filmmaker Alastair Cole understood from the beginning that his documentary Colours of the Alphabet on multilingualism in Zambia would need to have translation at its core, not least because it would not be viewed by any audience without subtitles or some form of translation, as is often the case in multilingual films. In order to adopt an accessible filmmaking approach, he identified two requirements: ‘the early engagement with the translation and subtitling process’ and ‘a full understanding of the ethical and representational issues that are imbued within it’, both of which were made possible thanks to the ‘unique translation context’ of observational documentary film (Cole 2015: 132).

As for the first requirement, instead of leaving subtitles for the end of the process, Cole regards them as ‘an intrinsic part of the filmmaking process that can be planned and engaged with from the start of production, and embraced as a powerful tool for emphasising perspective as well as forging characters and narratives’ (ibid: 134). Figure 1 shows how subtitling is integrated in Cole’s filmmaking workflow.  


Figure 1. Subtitling production process in The Colours of the Alphabet

Translation comes in before the editing process begins, when the rushes (unedited footage) spoken in Zambian languages are interpreted into English audio files. Selected sections of these files are then transcribed and subtitled as they are laid out on the editing timeline to form the first rough cut of the documentary, which includes proof reading and correction of the different languages used. Further editing work results in a fine cut of the documentary with English subtitles that have been spotted and fine-tuned according to guidelines. Interestingly, an Italian version of the subtitles for this fine cut is created before the film is locked, which allows to identify any particular translational issue that may be solved by modifying the editing of the film, such as the ones discussed in Section 5.3 with regard to Joining the Dots and A Grain of Sand. As explained by Cole, this is an essential step ‘to enable the subtitles to influence the pacing and emotional engagement of the film where necessary and to permit the adjustment of any scene that would create significant problems in viewing with the subtitles’ (ibid: 138). Once this is done, the film is locked and a subtitle guide is created by Cole and the producer of language and accessibility, including the full original transcript, the English and Italian subtitles and a section with notes on specific moments that may be challenging from a translational viewpoint. The subtitlers of the foreign versions are expected to communicate directly with Cole if need be.

The second requirement identified by Cole for the adoption of accessible filmmaking relates to the ethical implications of representing real people through translation. Unlike in fiction films, where translators deal with scripts that are usually prepared before production, (ethnographic) documentary filmmakers develop a direct and often personal relationship with their participants. The ethical implications of potential mistranslations are thus amplified, as the subtitles are an essential element in the accurate representation of those on screen. This raises very important questions regarding the filmmakers’ and translators’ responsibility in the representation of the people in the film:

The common, and often inevitable requirement to outsource the creation of various foreign language versions of films can result in removal of the director from any translation and subtitling debate, thus shifting ultimate responsibility for the translation and the representation of the characters involved in the film away from the person with whom the people in the film have entrusted their stories. Fully understanding the implications and procedures of subtitling, recognising the key role of the director in any debate, and understanding one’s own ideological perspective within the creation of the subtitles and the film as a whole is, I suggest, critical to mitigating the obvious dangers and harm mistranslation, and misrepresentation can entail. (ibid: 148)

In other words, it could be argued that the inclusion of translation and accessibility in the (pre-/post)production process is not only an issue of quality but also one of responsibility and trust for filmmakers towards their participants and their stories.

5.2. Translation and accessibility in (pre)production: mise-en-scene and cinematography

As briefly mentioned in Romero-Fresco (2013), the practical implementation of accessible filmmaking does not need to involve a radical change in filmmaking, but rather the consideration of a series of issues that are not normally addressed. The two aspects discussed in this section belong to the (pre)production stage and are costume (as an element of the mise-en-scene) and framing (as an element of cinematography).

A complex concept applied to the visual style in cinema, the mise-en-scene is described by Gibbs (2002: 1) as ‘the contents of the frame and the way they are organised’. For Bordwell and Thompson (1997/2007: 115), the mise-en-scene encompasses setting, costumes and makeup, lighting and staging. It is, for these authors, the most memorable aspect of filmmaking: the one that viewers remember more vividly and are more familiar with. The mise-en-scene allows directors to stage the event for the camera and essentially shows the director’s control over what happens in the film frame. As such, the mise-en-scene is also one of the elements that best epitomises the gap between film and translation. Filmmakers and film scholars have traditionally ignored the impact that the addition of letters at the bottom of the frame has on the visual style of a shot in which every aspect of the setting, costume, staging and lighting has been carefully considered. In some cases, the bright white or yellow letters of the subtitles may cause the completely opposite effect to the one intended by the dark lighting and the subdued tones used for a particular shot. The mise-en-scene will have changed dramatically, as will have the way in which the viewer watches the shot. However, none of that has been anticipated, designed, controlled or even seen by the filmmaker. It could thus be argued that the very defining feature of mise-en-scene (the director’s control over the visual arrangement of a shot), which in turn is the most memorable aspect of film for the viewers, is precisely what is lost in (industrial) subtitling, and it is what could be recovered with the more collaborative approach proposed in accessible filmmaking.

The mise-en-scene is also a key element in documentary filmmaking. Depending on where the documentary is placed in the above-mentioned continuum between ‘life caught unawares’ and ‘life arranged for the camera’, the mise-en-scene in this genre may refer to what is found in a given setting or to what is constructed for a given setting. In both cases, it is an essential aspect to provide information and influence the viewers’ mood, attitude towards the characters and, crucially, their belief in the situation (De Jong, Knudsen and Rothwell 2011: 155). Of the elements included by Bordwell and Thompson in the mise-en-scene (setting, costumes and makeup, lighting and staging), costumes are one over which documentary filmmakers may have control, as it is often possible to consult with the main participants what they will wear for an interview that has been arranged beforehand. Patterns, contrast, colours and shapes are the four areas to look out for here and the usual recommendation is to avoid whites, blacks, reds, oranges, high-contrast combinations and busy patterns, all of which can have a negative impact on the viewers’ experience (Nulph 2007). However, if filmmakers are planning to reach foreign or deaf audiences through subtitles and wish to have control over (or be aware of) the mise-en-scene, then they must also consider the impact that subtitles may have on it, and in this case the way in which the subtitles interact (or clash) with the setting and the costume. This lesson was learnt too late by the director of Joining the Dots, where the colour and pattern of Mag’s costume and Trevor’s coffee table clash with the black and white letters of the subtitles (see Figure 2 and Figure 3).


Figure 2. Mags’ interview in Joining the Dots


Figure 3. Trevor’s interview in Joining the Dots

In the case of A Grain of Sand, the clash is caused both by the patterns in Ofelia’s costume and the high exposure of part of the background setting, which was only partially solved by the filmmakers through the use of a black outline for the subtitles (see Figure 4).


Figure 4. Ofelia’s interview in A Grain of Sand

Another issue to be considered with regard to setting and costume is the potential use of colours in the subtitles, which is common practice in many countries in the case of subtitles for viewers with hearing loss. Given that the colours are assigned according to the on-screen presence and importance of every participant, both the filmmaker and the cinematographer can often make an educated guess as to what colour will be used for each participant and bear it in mind during the shoot. In other cases, though, the importance of the participants will be decided during the editing process. In the shot from A Grain of Sand shown in Figure 5, for example, yellow subtitles would cause an obvious clash.


Figure 5. Gloria’s interview in A Grain of Sand

In general, filmmakers who plan to use standard subtitles may be advised against using settings that are too cluttered at the bottom of the screen, as this will make it very difficult for the viewers to follow the dialogue.

Be that as it may, it follows that it is not only directors but also cinematographers who should be considering the impact that subtitles (if this is the translation/accessibility mode chosen) may have on their work. Whereas the filmmaker uses the mise-en-scene to control what is filmed, cinematography encompasses how the elements in the mise-en-scene are filmed (Bordwell and Thompson 2007: xiii). This includes the photographic image, duration of the image and framing. The latter, and especially the distance of framing or shot scale (i.e. the use of extreme long shots, long shots, medium shots or close-ups, among others) is particularly relevant for subtitling.

In order to illustrate the multiple conversations going on in the theatre between the actors, the audiodescriber and the audience before the play starts, the cinematographer of Joining the Dots opted for a long shot that includes fifteen people, while the sound editor allowed for some of their voices to overlap (see Figure 6). The image (the mise-en-scene and the shot size used by the cinematographer) and the sound tell the same story.  


Figure 6. Long shot of the theatre stage and audience in Joining the Dots

However, when the scene is subtitled (see Figure 7), the translator must choose who to translate, in this case the off-screen narration, which leaves out five other conversations that are going on simultaneously in the scene[3].


Figure 7. Long shot of the theatre stage and audience with subtitles in Joining the Dots

Although displayed together, the film and the subtitle are clashing head-on and telling a different story. The film is in ‘long-shot mode’ (overlapping voices by a group of people) and the subtitle is, as is normally the case, in ‘close-up mode’ (one voice or two voices shown consecutively). Needless to say, both deaf and hearing viewers can still see the group of people and hearing viewers can also hear the overlapping foreign voices, but the fact remains that standard subtitles cannot help but ‘single out subjects’ (MacDougall 1998: 169) and are often the verbal equivalent of a visual close-up. In scenes such as this one, and unless a different type of subtitle is used (see integrated titles in Section 5.3), subtitle viewers must live with a clash that does not exist for the viewers of the original film.

Another relevant issue regarding framing, perhaps the most common one in the interaction between cinematography and subtitles, is the use of close-ups. This shot has become increasingly common over the decades (Salt 1999) and is now considered as one of the defining elements of the ‘intensified continuity style’ applied in modern filmmaking (Bordwell 2002). Close-ups, as other shots, are often framed following the rule of thirds, an imaginary grid made up of two horizontal and two vertical lines that has been used for harmonious composition in film, photography and painting since it was first coined by John Thomas Smith in 1797 in his book Remarks on Rural Scenery. According to this rule, important elements in the frame are normally placed at the intersection of the lines, called sweet spots (Mercado 2010: 7). In close-ups the participant’s eyes are often placed around a sweet spot, as is the mouth, which leaves little or no space below the chin for the subtitles to be displayed. As can be seen in Figure 8 and Figure 9, whereas the original viewers of the documentary Requiem for the American Dream (2015) watch a close-up framed as per canonical rules, foreign and deaf viewers are presented with a shot that is not only cluttered but also prevents them from looking at the mouth of the speaker, which for hard-of-hearing viewers is essential to understand the dialogue.


Figure 8. Close-up of Noah Chomsky in Requiem for the American Dream


Figure 9. Subtitled close-up of Noah Chomsky in Requiem for the American Dream

As shown in Figure 10 and Figure 11, the director and cinematographer of A Grain of Sand took this into account when framing the main interviews of the film, which adhere to the rule of thirds. The interviewees’ eyes are around one of the sweet spots and in this case there is enough space to display the subtitles at the bottom without covering their mouths.


Figure 10. Medium close-up of Sady in A Grain of Sand


Figure 11. Medium close-up of Juan Carlos Nemocón in A Grain of Sand

As a matter of fact, the use of a bigger shot size enabled the filmmakers to use the setting for characterisation purposes, thus complementing the information obtained by the costume. The viewer can see the gang member wearing a tracksuit and speaking outdoors, in front of a fenced yard, whereas the mayor is wearing a suit and speaking in his office in front of institutional flags. Needless to say, this does not mean that close-ups cannot be used along with subtitles, but rather that it may be useful for filmmakers (and cinematographers) to bear this in mind as they shoot. A way of addressing this issue would be for the cameraperson to have an indication through the viewfinder of where the subtitles would potentially be displayed, so that a decision can be made as to whether a particular shot is to be framed in a subtitle-friendly manner or not.

5.3. Translation and accessibility in postproduction: editing and integrated titles

The previous section dealt with the relationship between translation and some of the elements used in the frame (costume and setting as part of the the mise-en-scene), including how they are filmed (framing as part of cinematography); in other words, the organisation of space in the production stage. This section focuses on the post-production stage and more specifically on the use of transition shots/cutaways and overlaps in editing (the organisation of time) and the production of integrated titles.

Cutaways are part of the coverage obtained during the production, for instance, of an interview. They are used by editors to cut away from the main interviewee, show something else, and then cut back to the interviewee at a different time without jump cutting. Rabiger (2004) distinguishes between cutaways (a shot of something outside the frame) and an insert (an enlargement of something in the main frame). For this author, these shots, ‘drawing the viewer’s eye to significant detail, arise from (and are motivated by) the storyteller’s narrative intentions’ (ibid: 460). Similarly, transition shots allow editors to pivot from one sequence to the next and to link separate scenes (Shook et. al. 2000: 55). How these shots are edited has a very significant impact on the foreign, deaf and blind audiences watching the film. By using transition shots with no narration or interview over them, filmmakers can manage to bridge the gap between original and foreign viewers, as the translated film becomes, if only briefly, the same as the original one. These shots allow subtitle users to watch and not read, dubbing viewers to enjoy the image with no distraction from dubbing synchronies and blind audiences to listen to the description of the images. This is the case in the shot from A Grain of Sand shown in Figure 12, where every pair of shoes hanging from the wires stands for a dead person in the Colombian armed conflict. Original and foreign/deaf viewers are in a similar position to explore the shot and its meaning.


Figure 12. Shoes hanging from wires in A Grain of Sand

For blind viewers to have a similar experience to original and foreign/deaf viewers, the shot would need to be displayed for longer, so that they can listen to the description of the image and then have some seconds to process it and think about it. If the filmmaker and/or the editor decide to use dialogue or narration over a cutaway or a transition shot, they should leave the shot on screen for long enough so viewers can read the resulting subtitle and view the image. The cutaway in Figure 13 creates a contrast between what the interviewee is saying (lack of solidarity in the community) and the image of two children helping each other to climb down a hill. The duration of the shot and the short length of the utterance, subtitled in only one line, gives the viewer enough time to read and watch.


Figure 13. Two children helping each other down a hill in A Grain of Sand

Finally, filmmakers should be particularly careful with the use of cutaways and transition shots that combine narration or interview and on-screen text, as they are difficult to process for target viewers. The target version of Figure 14 would need a subtitle to translate the interview that is being heard in this shot and another one to translate the content of the on-screen signs (‘con paz haremos más’), which will make the resulting image as difficult to process as cluttered and aesthetically different from the original.


Figure 14. Posters on the wall in A Grain of Sand

The clash between subtitles and on-screen captions with the participants’ names, explored by Fox (2016b), is particularly relevant in documentaries, since the use of captions is one of the defining features of this genre. Aesthetic considerations will be discussed below but, at any rate, extra time should be allowed for this type of shot. In the example shown in Figure 15, Joan’s words have been edited down to one subtitle line so that the viewers can have enough time to read both the caption and the subtitle.


Figure 15. Joan Greening’s interview in Joining the Dots

Another important element in the editing regarding translation is the use of overlaps, also known as overlap cuts or split edits (Salt 2002). This technique is used to bring the sound earlier than the images (J-cut) or the images earlier than the sound (L-cut) in order to avoid straight cuts and thus soften the transition between shots (Rabiger 2004: 231). In a J-cut, for example, the editor uses anticipatory sound or dialogue to draw our attention to the next shot. Just how many frames before the next shot this dialogue is brought forward depends on many factors, none of which has so far been related to translation. However, subtitling guidelines do refer to these types of cuts, as they often cause subtitles to run over shots (BBC 2016, Netflix 2016). The recommendation is that for dialogue starting more than 12 frames before the shot change, the subtitle should come in before this shot change, whereas for dialogue starting less than 12 frames before the shot change the subtitle should be displayed in the next shot. The former carries an additional risk: the subtitle will be shown across cuts and may be read twice, once in each shot. This is thus one more consideration to be taken into account by filmmakers and editors when they are preparing a J-cut or any other type of split edit.

Finally, one of the most effective devices to implement accessible filmmaking is the use of integrated titles (Fox 2016a) or creative subtitles (McClarty 2012). These subtitles respond to the ‘specific qualities of the individual film text, giving the creative subtitler more freedom to create an aesthetic that matches that of the source text’ (ibid: 139) instead of being constrained by standard font types, sizes and positions. In short, they are ‘an aesthetic extension of the film itself’ (ibid: 149). The integrated titles produced for Notes on Blindness seek to achieve this through the use of four devices: the choice of typeface, the play with rhythm and music, the interaction with the image and their position as part of the mise-en-scene.

Very little research has been conducted in AVT regarding the meaning-making and aesthetic role played by typeface in subtitling. Despite the evidence from other fields that typeface always carries meaning (Nørgaard 2009: 158) and that design is also information (Van Leeuwen 2009: 29), typeface has been as invisible in translation as translation has been in film. However, in an integrated and collaborative approach such as accessible filmmaking, it makes sense to consider how the typeface of the subtitles can contribute to the typographical identity of the film (Fox 2016b). For Notes on Blindness, after a discussion with the filmmakers, the typeface chosen was Adelle. This is a serif typeface more commonly used in print than in subtitles, where sans serif typefaces such as Arial, Verdana or Helvetica are normally the standard choices. The aim was to portray in the subtitle typeface the literary look that permeates the script and especially Jon Hull’s narration (see Figure 16). In contrast, a standard, sans serif Roboto was used for the sound coming from the TV or the radio, with an innovative use of the superscript to indicate the source of the sound (TV, radio, cassette, etc.) for viewers with hearing loss (see Figure 17).


Figure 16. John recording in cassette in Notes on Blindness


Figure 17. John’s son watching TV in Notes on Blindness

As pointed out above, documentaries are particularly prone to using on-screen captions with the participants’ names and subtitles. As well as discussing with editors the amount of time needed for both the captions and the subtitles to be read by the viewers, translators would also need to liaise with the graphic designers in charge of the on-screen titles. Both professionals are responsible for using text on screen and yet their lack of communication means that all too often the graphic designers’ careful choice of a typeface that is in line with the visual identity of the film is destroyed by the conventional look of the subtitles. Figure 18 shows the different typographical options considered for the use of on-screen captions and subtitles in the documentary Out of Kibera (2012).


Figure 18. Different options for captions and subtitles in Ryan’s interview in Out of Kibera

In this example, the translator and the graphic designer considered different options to increase the legibility of the subtitle over a white background (the use of a black outline, the use of a banner, etc.) and different typefaces for the speaker’s captions. As pointed out by O’Sullivan (2016), much work needs to be done in AVT to find out what meaning is carried by every typeface, how this meaning is conveyed and especially the extent to which ‘typefaces and layouts have similar connotations in different languages’, in other words, an intercultural grammar of the typographic mode (ibid.). In more practical terms, it is important for translators and graphic designers to start working together so that the use (and look) of text on screen can be coordinated.

As well as using typography to be in line with the look and feel of the film, integrated titles resort to rhythm and music to blend in the soundtrack and to the interaction with images and unconventional positioning to blend with the visuals. Firstly, add-on subtitles are used for some conversations, which allows the viewers to read the dialogue as they hear it and prevents the subtitles from pre-emptying information, as shown in Figure 19 and Figure 20.


Figure 19. Use of add-on subtitles in Notes on Blindness (I)


Figure 20. Use of add-on subtitles in Notes on Blindness (II)

Add-ons are also used in A Grain of Sand to subtitle the main song for viewers with hearing loss (see Figure 21 and Figure 22). Given the importance of visual sound and rhythm for these viewers, a subtitle for every instrument used is displayed on the screen as the instrument makes its appearance on the soundtrack.


Figure 21. Title credits and music subtitles in A Grain of Sand


Figure 22. Opening image and music subtitles in A Grain of Sand

As shown in Figure 23, integrated titles also blend with image and sound by, for example, fading in and out in instances where either the image of the film is fading or the narration is particularly slow and reflective. This calls into question why subtitles should simply pop on and off the screen in a way that is sometimes incoherent with the tone conveyed by both the image and the sound.


Figure 23. Subtitle fading out in Notes on Blindness

This interaction with the image is sometimes achieved by combining spotting and film kinetics, in other words, by matching the disappearance of the subtitle (or in this case the on-screen caption) with the speaker’s hand movements, as in the example shown in Figure 24 and Figure 25.


 Figure 24. Interaction between Ofelia and the captions (I)


Figure 25. Interaction between Ofelia and the captions (II)

Finally, perhaps the most distinctive feature of integrated titles is their placement, which is based on a set of criteria established by Fox (see Wendy Fox in this issue) regarding speaker position, speaking direction, contrast, framing and presence of primary and secondary areas of focus. The aim is to ensure that the placement of the tiles is in line with the composition of the image while increasing legibility for the viewers. Indeed, if a shot has been framed so that the focus is placed on both sides of the frame, as in Figure 26, why should foreign and deaf viewers be forced to look down to the bottom centre of the frame, where nothing happens, thus wasting precious time to take in the details of the image?


Figure 26. Subtitle displacement/integration in Notes on Blindness

The first empirical reception study comparing standard and integrated titles in Joining the Dots (Fox 2016a) has shown promising results. The findings of the experiment show that while viewers take a little more time to find integrated titles than standard subtitles, the overall reading time is reduced. With integrated titles, the viewers have more time available to watch and explore the images and they show very similar eye-movement patterns to the viewers of the original film with no subtitles. Furthermore, these integrated titles seem to increase the viewers’ suspension of disbelief, producing an enhanced sense of presence in the world of the film. These findings suggest that accessible filmmaking may bring about a degree of similarity in the way in which films are received by the source and target audience that is not normally found in films that are translated at the distribution stage. However, much more research is needed to investigate the impact that an accessible filmmaking approach can have on filmmakers, translators and viewers of original and translated films.

6. Conclusion

120 years after the invention of cinema and almost a century after the first use of subtitling and dubbing in film, translators are still, despite their importance, amongst the least valued professionals in the film industry. Similarly, 80 years’ worth of research in film studies has found very little to say about translation, accessibility and the viewing experience of foreign, deaf and blind audiences. Against this background, and in line with recent attempts to bring together the theory and practice of film and translation, accessible filmmaking offers an alternative approach aiming to integrate translation and accessibility as part of the filmmaking process through collaboration between translators and filmmakers.

The focus has been placed here on documentary films and on the reasons why this genre can potentially be very suitable for this approach. Documentaries adopting a realistic and observational life-caught-unawares approach, such as many ethnographic films, are based on a relationship of trust between the filmmaker and the participants that makes the filmmaker’s involvement in the translation of the film an issue of ethics and basic responsibility. On the other hand, more innovative, life-as-a-construct documentaries are often based on a sense of creativity that can open the door to new approaches to translation and accessibility. Finally, many documentaries, regardless of their style, make significant use of on-screen text (which concerns both original and translated films) and are made outside the constraints of mainstream fictional films, thus allowing translators to interact with the creative team of the film.

It could be argued that the traditional inclination of AVT scholars and trainers to use mainstream fiction films for research and training has, in a way, helped to perpetuate the gap between film and translation, given the difficulties involved in interacting with the filmmakers and in manipulating the material due to copyright restrictions. Looking at the margins (Brown 2015), where some of these documentaries are made, may help scholars to explore new research possibilities and our students to discover that there is life outside the mainstream. It may also reveal new ways in which the creative treatment of actuality, which is the signature of this genre, can include foreign and sensory-impaired viewers from conception. Realistically, this approach will not replace the current industrialised model that keeps translation and accessibility as an afterthought in the film industry, but it can offer an alternative for those filmmakers who really wish to make film for all.


Barbash, Ilisa, and Lucien Taylor (1997) Cross-Cultural Filmmaking, Berkeley, LA, London, University of California Press.

Bordwell, David (2002) “Visual Style in Contemporary American Film”, Film Quarterly 55, no. 3: 16-28.

----, and Kristin Thompson (1997/2007) Film Art: An Introduction, New York, McGraw-Hill.

Bosseaux, Charlotte (2015) Dubbing, Film and Performance: Uncanny Encounters, Bern, Peter Lang.

British Broadcasting Corporation (BBC) (2016) “Subtitle Guidelines”, URL: http://bbc.github.io/subtitle-guidelines (accessed 1 December 2016).

Brown, William (2015) “Politicizing Eye tracking Studies of Film”, Refractory: A Journal of Entertainment Media 25, URL: http://refractory.unimelb.edu.au/2015/02/07/brown/ (accessed 1 December 2016).

Cattrysse, Patrick, and Yves Gambier (2008) “Screenwriting and translating screenplays” in The Didactics of Audiovisual Translation, Jorge Díaz Cintas (ed), Amsterdam/Philadelphia, John Benjamins: 39–57.

Chanan, Michael (2007) The Politics of Documentary, London, BFI.

Chaume, Frederic (2004) Cine y traducción, Madrid, Cátedra.

Cole, Alastair (2015) Good Morning, Grade One. Language ideologies and multilingualism within primary education in rural Zambia, PhD diss., University of Edinburgh, UK.

De Higes Andino, Irene (2014) Estudio descriptivo y comparativo de la traducción de filmes multilingües: el caso del cine británico de inmigración contemporáneo, PhD diss, Universitat Jaume I, Spain.

De Jong, Wilma, Erik Knudsen, and Jerry Rothwell (2011) Creative Documentary. Theory and Practice, Harlow, Longman.

Díaz-Cintas, Jorge (2004) “Subtitling: the long journey to academic acknowledgement”, Jostrans - The Journal of Specialised Translation 1: 50-70, URL: http://www.jostrans.org/issue01/art_diaz_cintas.php (accessed 30 May 2017).

Fadiman, Dorothy, and Tony Levelle (2008) Producing with Passion: Making Films That Change the World, California, Michael Wiese Productions.

Ferreira, Cláudia (2002) “A legendagem de filmes de divulgacao científica: Crıítica de traducao de um argumento” in A divulgacao científica nos media contributos, M.J. Faceira (coord), Avanca: Cine-clube de Avanca: 105-121.

Fox, Wendy (2016a) “Integrated Titles - An Improved Viewing Experience? A Contrastive Eye Tracking Study on Traditional Subtitles and Integrated Titles for Pablo Romero-Fresco’s Joining the Dots” in Eyetracking and Applied Linguistics, Silvia Hansen-Schirra and Sambor Grucza (eds), Open Access Book Series Translation and Natural Language Processing: 5-29.

---- (2016b) “Should she really be covered by her own subtitle?  Text elements in film and their graphical translation”, Translation Spaces 5, no. 2: 244-270.

Franco, Eliana (2001) “Voiced-over television documentaries. Terminological and conceptual issues for their research”, Target 13, no. 2: 289-304.

----, Anna Matamala, and Pilar Orero (2010) Voice-over translation: An overview, Bern, Peter Lang.

Gibbs, John (2002) Mise-en-scène: film style and interpretation (Short Cuts), London, Wallflower Press.

Grierson, John (1933) “The Documentary Producer”, Cinema Quarterly 2, no. 1: 7-9.

Hanoulle, Sabien, Véronique Hoste, and Aline Remael (2014) “The efficacy of terminology-extraction systems for the translation of documentaries”, Perspectives: studies in translatology: 359-374.

Henley, Paul (1996) “The Promise of Ethnographic Film”, Lecture given in honour of the late Professor Paul Stirling at the University of Kent at the 5th International Festival of Ethnographic Film.

---- (2010) The Adventure of the Real. Jean Rouch and the Craft of Ethnographic Cinema, Chicago, University of Chicago Press.

Izard, Natalia (2001) “Doblaje y subtitulación: una aproximación histórica” in La traducción para el doblaje y la subtitulación, Miguel Duro (ed), Madrid, Cátedra: 189–209.

Kaufmann, Francine (2008) “Le sous-titrage des documentaires: Défis et enjeux de l’établissement du texte de départ” in La traduction audiovisuelle. Approche interdisciplinaire du sous-titrage, Adriana Serban, & Jean-Marie Lavaur (eds), Bruxelles, De Boeck : 69-83.

Krämer, Mathias, and Eva Eppler (forthcoming) “The deliberate non-subtitling of L3s in a multilingual TV series: the example of Breaking Bad”, Meta.

Kruger, Jan-Louis, Stephen Doherty, and María T. Soto Sanfiel (2015) “Psychological immersion in a foreign language. The effect of same-language subtitles on English foreign language viewers”, paper presented at the Conference Media for All 6 on September 17th 2015.

Lambourne, Andrew (2012) “Climbing the Production Chain”, paper presented at 2012 Languages and The Media Conference (Berlin, 21–23 November 2012).

Lewis, E. Douglas (2004) Timothy Asch and Ethnographic Film, London, Routledge.

MacDougall, David (1998) Transcultural Cinema, Princeton, NJ, Princeton University Press.

MacDougall, David (2001-2002) “Colin Young, Ethnographic Film and the Film Culture of the 1960s”, Visual Anthropology Review 17, no. 2: 81-88.

Martínez Sierra, Juan José (2012) “On the Relevance of Script Writing Basics in Audiovisual Translation Practice and Training”, Cadernos de Tradução 29, no. 1: 145–163.

Matamala, Anna (2009) “Translating documentaries: From Neanderthals to the Supernanny”, Perspectives, 17, no. 2: 93–107.

---- (2010) “Terminological challenges in the translation of science documentaries: A case-study”, Across Languages and Cultures 11, no. 2: 255–272.

----, Pablo Romero-Fresco, and Lukasz Daniluk (2017) “The Use of Respeaking for the Transcription of Non-Fictional Genres. An Exploratory Study”, Intralinea. URL: http://www.intralinea.org/archive/article/2262

Mercado, Gustavo (2010) The Filmmaker's Eye: Learning (and Breaking) the Rules of Cinematic Composition, Amsterdam, Focal Press.

McClarty, Rebecca (2012) “Towards a Multidisciplinary Approach in Creative Subtitling” Monographs in Translating and Interpreting (MonTI) 4: 133–155.

Netflix (2016) “Timed Text Style Guide: General Requirements”, URL: https://backlothelp.netflix.com/hc/en-us/articles/215758617-Timed-Text-Style-Guide-General-Requirements (accessed 1 December 2016).

Nichols, Bill (2001) Introduction to Documentary, Indiana, Indiana University Press.

Nørgaard, Nina (2009) “The Semiotics of Typography in Literary Texts. A Multimodal Approach”, Orbis Litterarum 64, no. 2: 141-160.

Nornes, Abe. M. (2007) Cinema Babel. Translating Global Cinema, Minnesota, University of Minnesota Press.

Nulph, Robert G. (2007) “Just What Should I Wear?”, Videomaker Magazine, October 2007, URL: https://www.videomaker.com/article/c14/12990-just-what-should-i-wear (accessed 1 December 2016).

Orero, Pilar (2006) “Synchronization in voice-over” in A new spectrum of translation studies, José María Bravo (ed), Valladolid: Publicaciones de la Universidad de Valladolid: 255-264.

O’Sullivan, Carol (2011) Translating Popular Film, Basingstoke: Palgrave Macmillan.

---- (2016) “Text on screen: (sub)titling and the 'typographic image”, paper presented at the EST Congress, at Aarhus University on 15th September 2016.

Pérez González, Luis (2012) “Co-Creational Subtitling in the Digital Media: Transformative and Authorial Practices”, International Journal of Cultural Studies 16, no. 1: 3-21.

Rabiger, Michael (2004) Directing the Documentary, Amsterdam, Elsevier.

Redmond, Sean and Craig Batty (eds) (2015) “Seeing into Things: Eye Tracking the Moving Image”, Refractory: A Journal of Entertainment Media 25,

URL: http://refractory.unimelb.edu.au/2015/02/06/volume-25-2015/ (accessed 1 December 2016).

Remael, Aline (2004) “A place for film dialogue analysis in subtitling courses” in Topics in Audiovisual Translation, Pilar Orero (ed), Amsterdam/Philadelphia, John Benjamins: 103-126.

Romero-Fresco, Pablo (2013) “Accessible filmmaking: Joining the dots between audiovisual translation, accessibility and filmmaking”, Jostrans - The Journal of Specialised Translation 20: 201–223, URL: http://www.jostrans.org/issue20/art_romero.php (accessed 30 May 2017).

---- (2018) “Eye-tracking, Subtitling and Accessible Filmmaking” in Seeing into Screens: Eye Tracking and the Moving Image, Tessa Dwyer, Claire Perkins, Sean Redmond and Jodi Sita (eds), London, Bloomsbury.

Ruoff, Jeffrey (1994) “On the Trail of the Native's Point of View”, CVA Newsletter 2: 15-18, URL: https://www.dartmouth.edu/~jruoff/Articles/CVANewsletter.htm (accessed 1 December 2016).

Salt, Barry (2002) “How They Cut Dialogue Scenes”, Cinemetrics, URL: http://www.cinemetrics.lv/dev/cutdial.php (accessed 1 December 2016).

Sanz, Elena (2015) Beyond Monolingualism: a Descriptive and Multimodal Methodology for the Dubbing of Polyglot Films, PhD diss, University of Edinburgh, UK.

Shook, Fred Shook, John Larson and John DeTarsio (2000) Television Field Production and Reporting, Amsterdam, Focal Press.

Taylor, Richard (1998) (2nd ed.) Film Propaganda: Soviet Russia and Nazi Germany, London, New York, Tauris.

The Guardian (2004) “Cinema verité”, The Guardian (editorial), Friday 23 July 2004, URL: https://www.theguardian.com/news/2004/jul/23/leadersandreply.mainsection (accessed 1 December 2016).

Thomas Smith, John (1797) Remarks on Rural Scenery, London, Joseph Downes.

Trinh, T. Minh-Ha  (1992) Framer Framed, New York, Routledge.

Van Leeuwen, Theo (2009) Introducing Social Semiotics, London, Routledge.

Vertov, Dziga (1984) Kino-eye: The Writings of Dziga Vertov, Berkeley, LA, London, University of California Press.

Vincendeau, Ginette (1999) “Hollywood Babel: The Coming of Sound and the Multiple-Language Version” in “Film Europe” and “Film America”: Cinema, Commerce and Cultural Exchange 1920-1939, Andrew Higson and Richard Maltby (eds), Exeter, University of Exeter Press: 207–225.

Zhang, Jinghong (2012) “The Interaction between Visual and Written Ethnography in Subtitling”, Visual Anthropology: Published in cooperation with the Commission on Visual Anthropology 25:5: 439-449.


A Grain of Sand (2014), United Kingdom, Kate Dangerfield and Arianna Avruscio.

Avatar (2009), USA, James Cameron.

Babel (2009), USA, Alejandro Gómez Iñárritu.

Charlotte Gray (2001), USA, Gillian Armstrong.

Home Sweet Home (2012), United Kingdom/France, Enrica Colusso.

Inglourious Basterds (2009), USA, Quentin Tarantino.

Joining the Dots (2012), United Kingdom, Pablo Romero-Fresco.

Land without Bread (1933), Spain, Luis Buñuel.

Lone Star (1996), USA, John Sayles.

Man with the Movie Camera (1929), Russia, Dziga Vertov.

Men with Guns (1997), USA, John Sayles.

Moana (1926), USA, Robert Flaherty.

Mystery Train (1989), USA, Jim Jarmusch.

Nanook of the North (1922), USA, Robert Flaherty.

Night on Earth (1991), USA, Jim Jarmusch.

Nina, (2012), Italy, Elisa Fuksas.

Notes on Blindness (2016), United Kingdom, James Spinney and Pete Middleton.

Out of Kibera (2016), United Kingdom, Pablo Romero-Fresco.

Requiem for the American Dream (2015), USA, Kelly Nyks, Peter D. Hutchison, Jared P. Scott.

Secret City (2012), United Kingdom, Michael Chanan.

Sherlock (2010-), United Kingdom, Mark Gatiss and Steven Moffat.

Slumdog Millionaire (2008), United Kingdom, Danny Boyle.

The Colours of the Alphabet (2016), United Kingdom, Alastair Cole.

The Revenant (2009), USA, Alejandro Gómez Iñárritu.

Train Pulling into a Station (1985), France, Auguste and Louis Lumière.

Triumph of the Will (1934), Germany, Leni Riefenshtal.

Workers Leaving The Lumière Factory in Lyon (1985), France, Auguste and Louis Lumière.


[1] Unlike Nornes (2007) and McClarty (2012), Fox (2016) avoids the term ‘subtitle’ and refers instead to (integrated) titles, which highlights the fact that they are placed directly into the picture and not necessarily at the bottom of the screen.

[2] An interesting example of consideration (and eventual refusal) of subtitles in the production process can be found in O’Sullivan (2011: 28-29) with regard to Charlotte Gray (Armstrong, 2001).

[3] Even if the translator had chosen to subtitle two of those conversations, standard subtitles could only show them consecutively and not simultaneously.


About the author(s)

Pablo Romero Fresco is a Ramón y Cajal grantholder at Universidade de Vigo (Spain) and Honorary Professor of Translation and Filmmaking at the University of Roehampton (London, UK). He is the author of the book Subtitling through Speech Recognition: Respeaking (Routledge) and the editor of The Reception of Subtitles for the Deaf and Hard of Hearing in Europe (Peter Lang). He has collaborated with several governments, universities, companies and user associations around the world to introduce and improve access to live events for people with hearing loss around the world. He is the leader of the EU-funded projects “MAP: Media Accessibility Platform” and “ILSA: Interlingual Live Subtitling for Access” and of the international research hub OGAM (Galician Observatory for Media Access). Pablo is also a filmmaker and has developed the notion of accessible filmmaking, the integration of translation and accessibility as part of the filmmaking process. His first documentary, Joining the Dots (2012), was used by Netflix as well as schools around Europe to raise awareness about audiodescription.

Email: [please login or register to view author's email address]

©inTRAlinea & Pablo Romero Fresco (2017).
"Accessible Filmmaking in Documentaries"
inTRAlinea Special Issue: Building Bridges between Film Studies and Translation Studies
Edited by: Juan José Martínez Sierra & Beatriz Cerezo Merchán
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2251

A proposed set of placement strategies for integrated titles

Based on an analysis of existing strategies in commercial films

By Wendy Fox (pixelpublic GmbH, Germany)

Abstract & Keywords

This article gives an overview of creative subtitling of additional languages used in recent commercial films. Based on a quantitative analysis of placement strategies of these integrated titles, possible shortcomings and improvements are discussed and the most frequent strategies are summarised in a first set of basic placement strategies.

Keywords: audiovisual translation, integrated titles, creative subtitles, placement strategies

©inTRAlinea & Wendy Fox (2017).
"A proposed set of placement strategies for integrated titles Based on an analysis of existing strategies in commercial films"
inTRAlinea Special Issue: Building Bridges between Film Studies and Translation Studies
Edited by: Juan José Martínez Sierra & Beatriz Cerezo Merchán
This article can be freely reproduced under Creative Commons License.
Stable URL: http://www.intralinea.org/specials/article/2250

1. Introduction

While there is a wide range of thoughtful guidelines for the creation of subtitles in general (e.g. Ivarsson and Carroll 1998, Karamitroglou 1998, Díaz Cintas and Remael 2007) and subtitles for the deaf and hard-of-hearing in particular (SDH, e.g. Ford Williams 2009), some audiences and film professionals still see subtitles as ‘a blemish on the film screen’ (Díaz Cintas and Remael 2007: 82) and ‘intrusion into the visual space of film’ (Thompson 2000: 1). This subjective perception hints on the relevance of not only proper translation and timing of subtitles but also their design. The position and design of traditional subtitles can be seen as a major drawback, and especially viewers not used to them – mainly from English speaking countries and those with a long tradition of dubbed films – might find it hard to focus on both reading the subtitles and exploring the image. British director Danny Boyle, appreciated for films such as Trainspotting (UK 1996) and Slumdog Millionaire (UK/FR/USA 2008), described the problem by stating that ‘you don’t watch the film – you read the film and you scan occasionally to the actors’ (Beckman 2008). And he doesn’t stand alone with this opinion. Rawsthorn (2007) describes conventional subtitles as ‘limply at the bottom of the screen’ (ibid.) and only legible ‘if you’re lucky’ (ibid.). The demand for the cheapest possible solution that works on all kind of devices might be an understandable explanation – but how much might a slightly higher investment into subtitling actually weigh, especially compared to the enormous costs of dubbing? Combined with Romero-Fresco’s observation that ‘half of the revenue of […] both top-grossing and award-winning Hollywood films comes from foreign territories’ (Romero-Fresco 2013: 202) while ‘translation and accessibility services only account for 0.1 % – 1 % of the budget of an average film production (Lambourne 2012)’ (ibid.), it can only be in the interest of film producers to take a critical look at the position and design of the subtitles on the screen and the way they are perceived by the audiences to introduce possible improvements.

Based on my doctoral thesis that uses eye tracking and questionnaires to investigate the impact of integrated titles on the reception and enjoyment of film (Fox fc.), this article focuses on the placement strategies of commercially used integrated titles - titles that are placed individually and integrated into the image composition of a film -, including how they came into existence, and what the critics and audiences thought of them. Based on an analysis of the frequency of the various placement strategies, it aims to create a useful and frequency-based set of positions for integrated titles.

2. Integrated Titles

Various terms have emerged over past years to describe deviations from conventional subtitling. While fan-created subtitles that ignore existing conventions were initially described as ‘abusive’ (Nornes 1999: 18), Nornes appears to be one of the first to discuss subtitles not only based on content but also layout. He recognises the criticism concerning the graphical intrusion of subtitles into the film image and considers it ‘likely that no one ever has come away from a foreign film admiring the translation’ (1999: 17), emphasising the need for exploring new methods. More recently, Foerster criticises the conventional guidelines’ aim of invisibility and the resulting ‘register and […] design for subtitles that never call attention to themselves’ (2010: 82), defining subtitles ‘solely as a means of understanding what is being said on screen’ (ibid.). Mentioning examples such as Monty Python and the Holy Grail (UK 1975), Annie Hall (USA 1977), and ‘Desperanto’ (segment in Montreal Stories, CAN 1991) and discussing the subtitles of Nochnoy Dozor (‘Night Watch’, RUS 2004[1]), she defines ‘aesthetic subtitling’ (2010: 85) as a practice that ‘draws attention to the subtitles via aesthetic means exploring semiotic possibilities, which include the semantic dimension without being restricted by it’ (ibid.) and is ‘predominantly designed graphically to support or match the aesthetics of the audiovisual text and consequently develop an aesthetic of their own’ (ibid.).

Following a more film studies-based approach, McClarty speaks of ‘creative subtitles’ (2012: 138). She sees ‘subtitling practitioners [as] mere norm-obeying machines’ (2012: 135) that ‘continue to have their hands tied by the constraints of the field and the norms of the profession’ (ibid.) while failing to ‘acknowledge the insights that could be gained by referring to audiovisual translation’s parallel discipline: film studies’ (ibid.). McClarty sees the need of a ‘creative’ approach that does not simply ‘describe a subtitling practice that differs from the norm but [that denotes] an approach that looks outward from its own discipline as well as its own culture’ (ibid.) and aims for ‘difference rather than sameness’ (McClarty 2012: 140). She emphasises that, so far, not the creativity of a translator or subtitle practitioner have led to innovation in title design, e.g. in Slumdog Millionaire, La Antena (AR 2007) or Austin Powers in Goldmember (USA 2002), but the ‘imagination of film directors and editors’ (ibid.) that did not only want to provide a translation but also create additional effects of comedic or artistic nature (2012: 140-142) as well as indicate sound (McClarty 2012: 146) or speaker location (McClarty 2012: 148). The goal should not be to replace existing norms and guidelines but a ‘creative response to individual qualities within and between films’ (ibid.), created by a ‘translator-title designer’ (McClarty 2012: 149) that aims for both linguistically and aesthetically pleasing titles that are produced during the filmmaking process and in cooperation with the film maker, editors, and title designers. Other terms that can be found in recent studies are ‘dynamic subtitles’ (Armstrong/Brooks 2014, Brown et al. 2015) and ‘speaker-following subtitles’ (Hong et al. 2010; Hu et al. 2013). These concepts place titles close to speakers and additionally focus on the automation of the placement process.

While the presented terms are still based on the concept of subtitles that are placed below or at the bottom of the image, even Nornes back in 1999 put the ‘sub’ in parentheses ‘because they were not always at the bottom of the frame’ (Nornes 1999: 23). As Bayram and Bayraktar speak of ‘integrated formats’ (2012: 82) when integrating text into image for educational purposes, and Armstrong and Brooks mention the enhancement of subtitles through ‘integrating them with the moving image’ (2014), the term of ‘integrated titles’ was chosen for the two conducted reception studies (Fox 2012; Fox fc.). The goal was to include all the previously mentioned concepts while emphasising the relationship between image and title. The various terms, however, show the wide bandwidth of possible approaches that cannot be defined by tight norms or strict guidelines.

As Foerster (2010), McClarty (2012), and others stated, not only the analysis (cf. Chaume 2004) but also the creation of subtitles should not purely be based on translation studies. Film studies can offer insight into image composition and storytelling, communication and graphic design can offer definitions of aesthetics and creativity, usability studies provide basics of user experience design as well as interface design, and computer sciences provide insight into potential automation processes and software design (Fox fc.). Based on the discussed research and these additional fields, the following section provides an overview and quantitative analysis of integrated titles in recent film productions.

3. Placement Strategies

This analysis is product-based and was conducted by taking screenshots (of static, unmoving titles) and short clips (of animated or moving titles) from a selection of films. The sample for analysis consists of – to the investigator’s best knowledge – all examples of English films released between 2000 and 2015 that contain integrated titles that translate an additional language into English as well as the so far only known film that made use of integrated titles throughout the whole length of the film, translating Russian into English. Contrary to the sometimes partially integrated SDH with horizontal displacement that can be found on Blu-ray disks (BD), there are no pre-existing guidelines or rules these integrated titles follow concerning placement and design. Therefore, all known examples were included in the sample analysis to reflect the state-of-the-art as accurately as possible while being able to exclude rare strategies due to their lower frequency. After a discussion of possible shortcomings, the most frequent strategies are then summarised in a first basic set of placement strategies for the creation of integrated titles.

3.1 Sample Analysis

Apart from the Russian film Nochnoy Dozor, the multilingual films and series discussed in this section were all produced in the USA and include an additional language that was subtitled into English. As the number of films including this kind of titles is still very small, no distinction was made between the forms and sources of the films analysed. Table 1 represents the analysed sample and gives the basic film credits.





Production Studio(s)

Main Distributor


Man on Fire


Tony Scott

Warner Bros. International

Twentieth Century Fox


Nochnoy Dozor



Timur Bekmambetov

Channel One


Twentieth Century Fox



(Episode 01)


Tim Kring

Tailwind Productions

Universal Pictures


Slumdog Millionaire


Danny Boyle

Celador Films Ltd.

Twentieth Century Fox



Fast Five[3]


Justin Lin

Original Film

One Race Films, Dentsu

Universal Pictures



Star Trek Into Darkness


J.J. Abrams

Bad Robot

Paramount Pictures



John Wick


Chad Stahelski

David Leitch






Table 1 Overview of the analysed films with integrated titles for an additional language

Tony Scott’s Man on Fire might not be the first film ever to use a more creative approach on its subtitles for additional languages, but it seems to be the most frequently mentioned early example (Kofoed 2011; McClarty 2012; Romero-Fresco 2013), followed quickly by the Russian film Nochnoy Dozor with its integrated titles for the English-speaking audience. Together with Nochnoy Dozor’s sequel Dnevnoy Dozor[4] (‘Day Watch’, RUS 2006) and Slumdog Millionaire, this ‘quartet of high-profile releases from Twentieth Century Fox and its specialty-film unit Fox Searchlight has initiated a radical, popular re-conception of the subtitle’s status in its previous distinction between inter- and intra-lingual contexts – as graphically additive captions and as linguistic translation’ (Kofoed 2011). They offer titles that are a ‘central aspect of the filmic mise-en-scène’ (Kofoed 2011)[5] and underline the fact that ‘subtitles as a graphic presence [are] inseparable from its medium as film’ (Egoyan/Balfour 2004: 68). While Man on Fire includes 219 integrated titles that translate from Spanish into English (compared to 1332 subtitles if watched with English subtitles, therefore accounting for approximately 16.44 per cent of the film’s dialogue), the English subtitled version of Nochnoy Dozor offers a complete translation for the foreign English-speaking audience with integrated titles accounting for around 6.46 per cent of the overall subtitles. Therefore, two kinds of integrated titles can be identified here: Titles that are part of the original version of a film and translate an additional language (Man on Fire, Heroes, Slumdog Millionaire, Fast Five, Star Trek Into Darkness, John Wick), and titles that translate an entire film into another language for a foreign target audience (Nochnoy Dozor, Dnevnoy Dozor). These films are presented and discussed in the following.

3.1.1 Man on Fire

‘Subtitles are boring’ – this simple opinion stated by Tony Scott (Scott, interview[6]) led to a new perspective on subtitles after decades of lacking interest from directors and producers. Without knowing that he would establish a new trend, if not even a future paradigm shift, he wanted the subtitles to be conceived as a ‘kind of character in the scene’ (ibid.). Therefore, his American remake Man on Fire provides a variety of integrated and aesthetically more or less pleasing titles that, as already stated, account for approximately 16.44 per cent of the film's dialogue (219 of 1332 overall English subtitles). With its setting in Mexico, several words and phrases in the film are in Spanish. However, not only this spoken content is subtitled, but also some English statements – presumably to underline and intensify the corresponding content. An especially interesting aspect is the conscious placement in the mise-en-scène or ‘within the shot’s depth of field, such that characters and objects may move before the subtitles, obscuring them from the reader’s gaze’ (Kofoed 2011). In addition to several fading animations (see Figure 1), a wide variety of kinetic effects was used to underline the active role of the titles in Man on Fire. Concerning the layout, the titles are rendered mostly in Franklin Gothic[7] and well readable due to their sans-serif and bold character that provides a good contrast to film image. They are also ‘not confined to the top layer of the film, they have depth and perception’ (Vit 2005).


Figure 1 Additional fading effect in Man on Fire (01:03:18)

The reactions to Man on Fire were mixed. While some critics celebrated the innovative titles as ‘small, visual victories that add charisma and personality to commonly bland and uninspiring subtitles’ (Vit 2005), others such as film critic Brian Gallagher refer to them as an example of ‘how NOT to do subtitles’ (Gallagher 2004). While Man on Fire displays several aesthetic types of text integration, it is not always comprehensible on what basis the present effect was chosen. Some titles are displayed word by word, corresponding to the actual speed of speech, while others are displayed in a whole or line by line, flicker, re-appear after a scene cut or move through the image. A smaller, but more comprehensive selection of effects might have had a stronger and more thoughtful impact. These integrated titles, however, seem to only be part of the English original version of the film. For the German DVD and BD releases, the titles were removed and replaced with traditional subtitles.

3.1.2 Slumdog Millionaire

‘Rejoice – subtitles have been freed!’ With this exclamation, the Washington Post celebrated Slumdog Millionaire’s titles when the film came to the cinemas in 2008 (Beckman 2008). While the film was initially supposed to be completely in English, the British director Danny Boyle quickly realized that the child actors in India wouldn’t be able to deliver their lines credibly in English (ibid.). He therefore decided to let them speak Hindi and promised Warner Independent[8] ‘that the film would be even more exciting because of the subtitles’ (Beckman 2008; cf. Kofoed 2011). He ‘came up with the idea to have the subtitles look more like the dialogue in comic books, which float depending on where the characters are positioned’ (Beckman 2008). According to McClarty, this ‘works to keep the audience engaged with the plight of the young boys’ (2012: 143). The titles were also ‘placed within coloured caption boxes’ (Kofoed 2011)[9] and inspired not only by comic book captions but also the non-traditional subtitles of Nochnoy Dozor (Beckman 2008; Kofoed 2011). The coloured boxes ensure both ‘visibility’ (McClarty 2012: 143) and legibility against the changing background and slightly resemble compositions on posters and in comic books. While traditional subtitling conventions often suggest black boxes (e.g. Karamitroglou 1998), the ‘colour of the transparent box changes according to the colour scheme of the mise en scène’ (McClarty 2012: 143) in Slumdog Millionaire. The 202 integrated titles account for about 17.99 per cent of the film’s dialogue (out of approximately 1123 subtitles if watched with English subtitles).

The Washington Post’s film critic Rachel Beckman described the titles in Slumdog Millionaire as bouncing ‘around the screen in a rainbow of colors’ (Beckman 2008) and ‘stylish and splashy and original. They’re liberated’ (ibid). McClarty underlines the importance of these titles not only fulfilling ‘their linguistic, translational function’ (McClarty 2012: 143) but ‘also fulfilling an equally important aesthetic function through their colour and an affective function through their positioning in the heart of the on-screen action’ (ibid.). However, she notes that these titles rather allow for an ‘overall understanding of the situation and dialogue, rather than a word by word comprehension of each subtitle’ (ibid.) and also Romero-Fresco sees that ‘the translational role of these subtitles (and even their legibility) often takes a back seat to their affective use of colour and position to advance plot and character development’ (Romero-Fresco 2013: 210). This discussion of the titles in Slumdog Millionaire shows that decisions concerning the significance awarded to the translational function of the titles on the one hand and the aesthetic and affective function on the other have a strong influence on the final product and translators and designers involved in the postproduction have to work hand-in-hand to achieve a good result.

In addition to ignoring the layout instructions of the traditional guidelines, the titles also lack periods at the end of each sentence. This might lead to irritation of the audience as it might not be clear whether the sentence was finished or will continue in a following title. The typography is modest and makes use of a single typeface with soft edges and a legible stroke width. The combination of typeface, drop shadow and coloured box, however, might lead to a decrease in legibility. From examples such as shown in Figure 2, it appears as if the image composition was considered more important than indication of speaker or speech direction, but then there are also titles that disturb the image composition strongly even though better solutions would have been possible. While the placement of the titles could be improved by decreasing the distance to the speakers and focus points, the overall layout and especially the coloured boxes support the atmosphere and emotions in the respective scenes and most likely had a strong impact on the reception of Slumdog Millionaire.


Figure 2 Indication of speaker outside the frame in Slumdog Millionaire (00:34:49)

3.1.3 Heroes

The American television show Heroes was produced and broadcast by NBC between 2006 and 2010.[10] While almost all characters in Heroes speak English, the Japanese friends Hiro and Ando speak their mother tongue at the beginning of the first season and only learn English bit by bit due to their continuous stay in the United States. Thus, Hiro and Ando are the only characters in the first season to be subtitled. These titles were integrated into the image dynamically, seemingly on the basis of the placement of speech bubbles in comics as Hiro is a huge fan of them and draws all his knowledge of supernatural abilities and other characters from them (see Figure 3). The integrated titles in the first episode of the first season of Heroes make up about 13.56 per cent of the film’s dialogue, accounting for 96 out of 708 titles if watched with English subtitles.


Figure 3 Title placed in between speakers during a dialogue in Heroes S01E01 (00:14:12)

Overall, the first episode of the first season of Heroes showcases many interesting placement strategies and might include some of the first film material that was actually shot in a way that would provide space and sufficient contrast to allow for the title placement in the postproduction.

3.1.4 Fast Five

Fast Five was released to cinemas in 2011 as a joint production between Original Film, One Race Films, and Dentsu, distributed by Universal Pictures. Even though the screened version included 101 individually placed titles translating Portuguese conversations and statements into English, no mention of this can be found in critic’s reviews of the fifth part of the Fast & Furious series. The integrated titles make up 7.49 per cent of the subtitles that would be required to watch Fast Five completely subtitled in English (101 out of 1348 subtitles in total).


Figure 4 Title indicating speaking direction (Fast Five, 00:11:51)

As visible from Figure 4, the typeface is a bold white sans-serif with a slight shadow and creates a strong contrast to most backgrounds. In one case, single words are highlighted in italics, reflecting the speaker’s emphasis. All in all, the titles are usually placed close to the speaker or in speaking direction and allow for a close focus for the audience. Even though the titles weren’t mentioned in any major film critic’s review, the production studio and directors seemed content and continued the practice in the following sequels of the Fast & Furious franchise – Fast & Furious 6 and Fast & Furious 7, both distributed by Universal Pictures. As the titles in the sequels follow similar layout and placement strategies, these films were not included in the analysis to prevent biasing of the placement analysis towards this franchise.

3.1.5 Star Trek Into Darkness

After the immense success[11] o