Séance N° 8 « Recherche et traduction (II) »

Dans l’édition de traduction, comment garantir la qualité du texte ? Qu’est-ce qu’une « bonne traduction » en SHS ? Quelle est la place du secrétariat scientifique ou de l’éditeur dans l’établissement d’un texte scientifique ?

Mark Mellor, directeur de la société de traduction Cadenza
Chloe Morgane, secrétaire de rédaction chargée de l’édition anglaise des Annales HSC
Nicolas Barreyre, historien, membre du Comité de rédaction des Annales HSC

Mark Mellor

Elusive Objects of Translation

Theory inspires practice: they intersect at creative choice. But how much choice does a translator have? In faithfully (or slavishly?) reproducing the—sometimes inexplicable—excogitations of an author, the translator surely lacks the freedom choice requires. But a translator who theorises wonders about what there is to translate, what must be attended to in a text. This investigation of the objects of translation is one side of the translator’s freedom; another side is how the resulting discoveries come to be incised into the new work as it progresses.

In this, the translator is not merely celebrating human individuality—like a woodsman wielding an axe with a characteristic flourish. It is the nature of translation itself that elicits the translator’s ontological questioning of an original text. A translation is more than a functional object. An axe, once made, works off its debt to the world through the cutting of wood. A translation is doubly indebted—to the world (its readers, its native language and culture, its discipline) and to the work from which it springs or struggles free. In asking whether my new text can truly translate a prior text, it is not enough to identify its purpose, what it does in and of itself, what use it is, what good it does; I need to know what I’m working upon, my objects, my material. When I take an axe to a trunk to make firewood, no matter how violent the transformation, I don’t efface the wood itself. For a monoglot reader of the original work—for a monoglot author—the translator renders the text unreadable, unrecognizable. And the transformed text becomes for its new audience both a surrogate and the authentic expression of the author’s thought. Still, after all the hewing and sawing, the syntactic uncouplings and rehitchings, the onamasiological windfalls, everything that matters—we ardently believe—has been preserved. What, then, do I bear across intact to my translation? What, as a translator, must I take care, faithfully, painstakingly, at all costs, not to efface, not to transform?

Here’s an obvious answer: meaning. It is a very common understanding of translation—confirmed by some prestigious dictionaries—that meaning is exactly what translation conveys. For example, the Oxford Companion to the English language tells us that ‘Translation is the communication of the meaning of a source-language text by means of an equivalent target-language text’. But no text, nor any speech act, has the purpose of conveying meaning. If I tell you ‘Your cat is in the road’, I may, depending on the circumstances, be describing the scene outside my house, giving you a warning, or deriving a cruel enjoyment from deceiving you. My purpose is certainly not to communicate to you the meanings of the words ‘your’, ‘cat’, ‘is’, and so on. I would not use these words unless I were already sure that you understood their meaning. In warning you, my intention is not to convey to you the meaning of my warning, but to warn you. And when I tell you a joke, my intention is to amuse you. We might say that meaning is the medium whereby I achieve these acts of transmission, not what is achieved. And this meaning is common to all languages. There is not a French meaning, a Gujarati meaning, an Inuit meaning. I’m tempted to say that meaning is the bearer of words and that translation is done as if the opposite were true.

Even if translation were about the conveying of meaning, one would have to ask whether it is all the meanings of all the words that are to be translated, which would result in a chaos of mismatched polysemous fields. I must choose what not to translate, as much as what to translate: the translator’s work is also to carefully suppress meaning. On what principle? I cannot choose which ‘word meanings’ to translate unless I understand the import of the statement that contains them; I cannot understand the import, that is, whether the author is describing, warning, joking… unless I understand how the statement gestures towards the world. For my present purposes, it is this world-directedness—language’s power to harness meaning for acts of gesturing—that interests me.

Before going on, I should make it clear that I’m talking about translation in the humanities and social sciences, not literary translation. A work of literature may be regarded as a collection of falsehoods that can tell us important truths about the human condition. An essay in the humanities or social sciences, having no obvious recourse to such magic, is constrained, when it uses factual statements, to align them with the world via the epistemic standards peculiar to its discipline. Texts in the humanities and social sciences contain statements that may be true or false, that correspond to the world in testable ways. Factual inaccuracy (false alignments) could impair or even demolish a text’s credibility. But it would be a sterile piece of writing that consisted of nothing but verifiable truths.

Democritus famously declared that there is nothing but atoms falling through the void. Defending this minimalist ontology against the criticism that, in such a world, atoms would be unable to interact, that causal relations would thus be ruled out and that, contrary to considerable evidence, nothing could ever happen in the world, Lucretius, in The Nature of Things, uses the idea of a swerve, a turn, a twist, or ‘clinamen’ to rescue the theory.

Let’s imagine a text that consists only of factual statements. They are the atoms arranged in parallel lines, never converging, always homing in on the state of affairs in the world to which they owe their fidelity. Is such a text ever written?

Suppose the first sentence of our unswerving ‘Democritean’ text is ‘The median number of books read by a British adult in one year is 4’ and the second sentence is ‘The average British citizen watches 32 hours of television per week’. We already have an argument! The mere juxtaposition of the two sentences, allows them to ensnare each other with their tiny hooks, announces a way of seeing the world that is more than the sum of the truths they express. A ‘Democritean’ text—theoretically possible only in a world like that of the Tractatus Logico-Philosophicus—could never convey an argument. And no such text is ever to be found in the humanities and social sciences, because a paper or a book is not a collection of facts, but an attempt to have us see the world in a certain way, a reconceptualization that uses a panoply of devices to achieve itself. The gravitational pull of the creative imagination, dense with value, bends our logic out of its straight paths and yields crooked truths. And the ways these align with the world are among the objects of translation.

Formal deductive procedures, inductive arguments, informal implicature, and all that rhetoric offers by way of persuasion—all of these align with the world. For it is not only factual, verifiable statements that point. Any critical reading of a text reveals the specific ways statements are being aligned with the world, whether they are constative, persuasive, admonitory, jocular—and therefore the epistemic standards to be applied to them. It is not only their fidelity or truth that is at stake—but also their use, their rootedness in, to borrow the later Wittgenstein’s notion, our human forms of life.

There are things in the world that make our statements about it true. That your cat is sitting in the road makes the statement ‘Your cat is sitting in the road’ true. And there are things about the world that allow us to grasp the difference between a warning and a straight description—the roaring engine of a car may help us here. Aside from factual statements, epistemic criteria also apply to how we should construe (how seriously, for example) promises, warnings, conjectures, even jokes. A promise may be believable or not, depending on who makes it, and in what circumstances; a warning may be insufficiently grounded in causes for alarm.

Unlike Lucretius’ clinamen, the swerve does not occur at random, but is interwoven throughout a text. Not beyond, but enchased in, logical strategy and inlaid in rhetorical devices of persuasion and dissuasion, the swerve awaits the attention of the translator. Texts and speech acts direct us to look at some part of the world in a particular way. They are gestures that it is the task of translation to replicate, not to perform, for it is another limitation of the translator’s creativity that such gestures are simulacra. But in identifying and re-presenting clinamina, the translator begins to come into her freedom.

Here are the first two sentences of Camus’ Le Mythe de Sisyphe: ‘Les dieux avaient condamné Sisyphe a rouler sans cesse un rocher jusqu’au sommet d’une montagne d’ou la pierre retombait par son propres poids. Ils avaient pensé avec quelque raison qu’il n’est pas de punition plus terrible que le travail inutile et sans espoir.’ The final words of the essay are these: ‘La lutte vers les sommets suffit à remplir un coeur d’homme. Il faut imaginer Sisyphe heureux.

Camus did not follow a straight path to contrive this fabulous reversal. The translator’s task is to discover it.


When we are led away from the idea that translation is about the conveying of meaning, it becomes clear what is at stake in saying something cannot be translated: we thereby denude all other languages of a gesture towards the world, a way of thinking it. Such a claim cuts too deeply into the possibility of human experience. I prefer to think of translation, whatever else it might be, as the ultimate trial of a text’s coherence. Its ability to detect and perpetuate the alignments that link an original text with the world reassures us that our thoughts make sense and that nothing need be consigned to silence.

A translation must pick up all alignments, must at least reproduce the reconceptualising swerves of an author’s argument. A well-written source text seldom puzzles a competent translator. But sometimes—and these are her most creative moments—a translator, having gathered the sheaf of multiple alignments that is a complex thought, holds it whole and steady in her mind… and does not know how to go on: it is her own language, within which there is no obvious way for the thought to be inscribed, before which she stands at a loss.

Are there objects of translation even more elusive—so indeterminate one might think they are not the business of translation at all? In any creative effort there are choices beyond reason. Kierkegaard said that the instant of decision was madness. Choices are only made when there is no clear rational reason for preferring one option over another. We see these moments in those creative, but not always deliberated juxtapositions of statements, the opting for this rather than that metaphor, an intertextual allusion, the momentous but undeliberated omission—the unaligned and uninscribed import of all writing and utterance.

The traces of these moments are there in the text—not inscribed as signs, but still worthy objects of the translator’s care and attention, and thus translatable. As are the prodigious misalignments of psychotic speech and writing, which by hook or by crook cling stubbornly to the quilting points of a common language and its inalienable rootedness in our world. When Descartes hypothesized that he might be mad, he still believed this did not invalidate the cogito. ‘Je pense, donc je suis’ is valid, we might say, even when I’m ‘not myself’. No, not even madness has ever been outside, but only at the extremes of human discourse—at the extremes and in the interstices, waiting to be found by translation.

Mark Mellor is editor in chief at Cadenza Academic Translations




Chloé Morgane

Le roman d’Albert Camus L’étranger a reçu en traduction anglaise deux titres : « The Outsider » et « The Stranger ». Cette réalité est le résultat d’une erreur de communication entre les deux coéditeurs américain et britannique dans la dernière phase de la préparation de  la publication du texte traduit en 1946. Cette anecdote rappelle combien la publication des traductions est tributaire d’un ensemble de questions pratiques, logistiques et financières, persistantes encore aujourd’hui à l’heure des communications instantanées via internet.

Depuis quelques années, la revue Annales histoire et sciences sociales s’est lancée dans la publication d’une version anglaise et depuis 2017, la revue est coéditée par les éditions de l’EHESS et Cambridge University Press. Les deux versions ne sont pas tout à fait similaires car les comptes rendus ne sont pas traduits en anglais et la publication anglaise est préparée une fois l’édition française achevée. En amont, il est demandé aux auteurs de rédiger une fiche de traduction susceptible d’aider le traducteur dans son travail, notamment en matière de terminologie, puis ce dernier a six semaines pour rendre sa première version. Le secrétariat de rédaction de la revue effectue en parallèle un travail de recherche sur les concepts, les noms propres, la translitération et la recherche de traductions canoniques. Une triple relecture du texte est ensuite effectuée puis les textes sont renvoyés aux auteurs. Il y a donc une communauté d’acteurs (traducteur, éditeur, expert, auteur) travaillant sur le texte traduit, entre lesquels les échanges autour de la traduction se font en transparence.

Quels types de difficultés rencontrons-nous ? Il y a d’abord des difficultés à trouver la traduction adéquate de certains concepts, ainsi le titre « La nomenclature socio-professionnelle : une histoire revisitée », qui sonne « très français » et que nous avons finalement traduit par « revisiting the history of socio-professional classification ». Autre exemple : nous avons décidé de maintenir au pluriel l’expression « statuts sociaux » qui est généralement utilisée au singulier en anglais depuis Max Weber [Social status], parce qu’il s’agit de décrire un contexte français et que cela fait sens ainsi.

Comme il s’agit d’un processus d’exportation de la revue -du français à l’anglais-, nous avons fait le choix de conserver les traces de la traduction quand elle est porteuse de sens, ce qui est une pratique assez nouvelle. Ainsi dans un article, nous avons décidé d’utiliser le mot « cadre » à cause d’un livre qui a utilisé ce gallicisme (Luc Boltanski « The making of the classe : cadres in French Society »).

En conclusion, on peut dire que dans ce processus très particulier de l’édition d’une version anglaise d’une revue scientifique en sciences humaines, une bonne traduction est une traduction « relue ».


Nicolas Barreyre

Le projet des Annales en anglais a débuté en 2012 et les premiers numéros ont été publiés en 2014. Aujourd’hui les épreuves sont propres, mais c’est le résultat d’un long processus qui a mobilisée toute une équipe. L’objectif était de faire une traduction ET une nouvelle édition. Le travail éditorial est un travail scientifique, la version de référence restant la version française.

Nous sommes partis du projet de faire une revue internationale française en anglais, mais nous ne voulions pas juxtaposer les articles en français et en anglais. De plus, les auteurs fréquentent une littérature de référence anglophone. Souvent, quand ils relisent leur texte traduit en anglais, ils affirment qu’ils auraient écrit autrement s’ils avaient écrit directement en anglais. C’est cet écart -parfois frustrant pour les auteurs- qui, pour nous, est important. C’est dans cette tension que se trouve une bonne traduction. Il s’agit d’un écart, d’une tension entre des habitudes historiographiques différentes.

Dans ce processus, la traduction est une épreuve pour les « belles plumes ». Pour notre part, nous n’avons pas peur des répétitions, la pratique du traducteur de textes de sciences humaines et sociales est en cela très différente de celle du traducteur littéraire.  Être clair ne signifie pas « lisser ».

La bonne traduction est la plus lisible, mais elle est aussi celle qui rend visible cet écart de la traduction, par exemple dans l’usage de certains mots, dans l’ajout de certaines notes explicatives, parce qu’en français, il y a beaucoup d’implicites. Notre question d’éditeur scientifique est alors de trouver jusqu’à quel point il faut que nous rendions l’implicite explicite.

Il ne s’agissait pas de faire une revue anglophone, anglaise ou américaine mais bien de continuer à être une revue française internationale. Le rapport à la vérité se construit dans le cadre éditorial de la revue, c’est pourquoi nous assumons certains choix de traduction.