Translating the later books of Cassius Dio's Roman History into French: a few methodological considerations

Traduit de :
Traduire les livres tardifs de l’Histoire Romaine de Cassius Dion en français : quelques questions de méthode



Propos de traducteurs



It has now been three years since I started my doctoral thesis titled Scholarly edition, translation and commentary of Cassius Dio's Roman History, books 75-771 (Édition critique, traduction et commentaire des livre 75 à 77 de l'Histoire Romaine de Cassius Dion). Most of this time was spent establishing the text itself2: reviewing manuscripts, editions, variants and fragments, a task made all the more difficult by the COVID 19 pandemic, which denied me access to my sources for quite a long time. For a time, then, I worked without giving thought to the second part of my thesis, translating a rough draft in French of Ursul Philip Boissevain’s edition of Dion (BOISSEVAIN 1901). This translation was to serve as a baseline of knowledge about the text. This, however, was done naively, and too quickly, and when the time came to make changes and corrections to my translation, I was ashamed for the sheer number of errors and misreadings which I inflicted upon Valérie Visa-Ondarçuhu, my director, and Éric Foulon, my former professor, who had both accepted to proofread my first attempts. The truth is that I had applied inapt methods to my doctoral research, as I often have these last three years. The methods which had earned me a masters degree were not up to par with this new task.

This is the question I now want to address: what are the differences between translating, as a student, relatively short extracts of literary works chosen by a teacher, as we have all done, and translating a text such as Roman History, a longer text, in an ancient language, and that we had to establish ourselves from manuscript sources ? I will first make salient some important features of the translation of literature from ancient languages to current ones, as there are some key differences between this and translation between living languages. Then I will compare translation as it was taught to me in French higher education with how I ought to translate, now that I am not exactly a student any more. Finally, I will expose some specific difficulties I encountered while translating Cassius Dio's Roman History, due to the fragmentary nature of the text.

1. Translating from a dead language into a modern one: a few thoughts

One of the most surprising things I discovered when I started researching for this paper was the relative scarcity of bibliographical resources concerning the methodology of translation from dead languages, at least from Latin and Greek. Most of the references I found consisted in translation commentaries on a specific author's corpus, an approach Philippe Heuzé recently used on Vergil (HEUZÉ, 2017), or commentaries on fine points in ancient translations between Latin and Greek, such as that undertaken in my masters degree (THÉROND-DEBAT, 2016), and those of the more experienced scholars (MASON, 1974; FREYBURGER, 1997) to whom I referred extensively. Few papers or monographs have focused upon the translation of ancient texts and its specificity in a general manner. The largest body of thought on the subject is to be found in manuals preparing students for the French teacher recruitment examinations, the CAPES and the Agrégation (LACAZE, 1999, 7-20)3. This is surprising, for although translation studies is a relatively recent field of research, it has been a fertile one, and it has been the source of many epistemological and methodological debates (ALBRECHT & MÉTRICH, 2016).

Apart from these few references, I was, at first, compelled to turn most of my research efforts to French translation studies concerned with living languages. For the most part, information pertaining to translation studies written in other languages did not retain my attention. Since my main objective is the translation of Ancient Greek into French, I directed my attention first to the efforts of French-speaking translators or scholars, and how they named and navigated between theories and norms in translation. This phase of my research was quite enriching, and it brought light to a particularity of ancient languages. They are, for the most part, also called “dead languages”, which means that, while practiced, they do not have any native speakers anymore. This also implies that there is no more new literature made in these languages4, which changes the very core of the relationship between ancient texts and their modern day translation. Why ? One could advance, as George Steiner does, that “When we read or hear any language-statement from the past, be it Leviticus or last year’s best seller, we translate” (STEINER, 1992, 28); that an operation of translation is always present with texts written in the past, however distant or recent, and thus that whatever the language of the source is, their interpretation is a form of translation. According to Steiner, therefore, there is no difference in nature between translating the latest book by Brandon Sanderson into French, doing the same with the Iliad, or interpreting in contemporary French the original XIXth century French of Hugo's Les Misérables. To Steiner, 1- translation is hermeneutic and every interpretation is in a way a translation, and 2- every translation is a displacement of the past into the present (STEINER 1998, 453-454).

However, while the hypothesis that every interpretation or translation involves the same intellectual processes is a tempting one, or can at least be argued, we have to take into account the sheer diversity in the types or methods of translation (is it an XVIIth century « Belle infidèle6 », a short extract translated as part of an exercise, or a modern scholarly edition?) and the differences in the type of activity involved (I am not undertaking the same action when I set myself to translating Cassius Dio's Roman History as when I interpret a colleague's words during a conference, no matter how complex their expression is). This diversity can be explained, to a certain extent, using Steiner's hermeneutic approach (different interpretations of a text yield different results), but this is not able to completely account for it, nor do we find in this approach a precise methodology in relation to the translation of ancient texts. It does not go further than allowing that ancient texts tend to have ancient-sounding translations (STEINER, 1998, 464). The fact is that both the greatest strength and the biggest weakness of Steiner's thesis is that, in attempting to write a general philosophy of translation, it ends up hypothesizing on the characteristics of an ideal practice of the art of translation, independently of context. In this attempt, when commenting on other translator’s work, he systematically takes the literary commentary of the text in its source language as his starting point, judging the translation from his own hermeneusis of the original text. Although Steiner’s literary analyses are subtle, this tendency may lead him to neglect, in the articulation of his arguments, two things that come before any hermeneutic work, perhaps even before the reading of the text: the precise relationship, both linguistic and cultural, between the two languages involved7, and the translator’s objectives in undertaking a translation. Both of these points will be developed in relation to ancient languages.

Let us begin with the attitude we have, at least in France, towards ancient languages and cultures. Modern translation studies have drawn attention to the importance of the relationship between translated languages beyond the scope of simple linguistic matters. In reality, relationships between cultures can be just as important. For example, in the domain of living language translation, researchers have noticed that translations were not devoid of power relations. Thus, in post-colonial contexts, translations of works of formerly colonized peoples can be done in such a way that the interests of the target system (the main readership and the market formed by former colonizers) are better taken into account than those of the source system (D’HULST, 2007). To put it differently, the way in which the translator views the source language (and the source culture) has an influence on their work. In this light the question becomes how we view ancient cultures, and how this affects the way we translate their literature. If we had to agree on a single adjective to describe our attitude towards ancient texts, it might be « deferential ». We see them (or at least some of them) as the foundations on which we built our own literature, and consider the languages in which they are written as the respected elders of our own. This deference, this positive social valorization of ancient cultures is one among many factors of intellectual interest for the Classics, including in the academic field and is one of the reasons why writers keep writing intrigues based upon the myths of no longer practiced religions8. It is also what explains Steiner's remark on ancient-sounding translations (STEINER, 1998, 464). Indeed, we do not end up translating this way only because of the vast amount of time elapsed between Plato writing the Gorgias and its translation into French, but also because we tend to associate antiquity with respectability. We translators of the Ancients therefore euphemize, in some cases, Plato’s lively dialogues or Aristophanes’ dirty jokes9, because the author's language, which would have sounded crude in Ancient Greek ears, cannot be allowed to be heard this way in French, out of respect as much as because of linguistic or cultural differences: there exists a « deference bias » in ancient text translation.

This phenomenon does not exist as much for living languages. One may sometimes recognize similar euphemizations in translations of works written in living languages. However, these can be fringe cases, regarding long established works whose language has had time to fall behind current convention, even in their original language. French translators treat Shakespeare with due deference, but the Bard's English does sound ancient to native English speakers' ears; that translations of his work would sound the same is therefore no call for surprise. There is, however, a key difference between Shakespeare and, for example, Cicero: an ancient-sounding translation of Shakespeare can aim to re-create for a foreign public the relative strangeness of his language to a modern native speaker; but an ancient-sounding translation of Cicero cannot replicate that, as there are no native Latin speakers. The strangeness of the Latin is not the strangeness of the same language after the passage of time: it is that of an entirely different, although related language, that we feel deferent towards. It could be argued that much of the same can be said of the differences between ancient and modern Greek, and that they are past the point where they can be considered to actually be the same language10.

2. Learning to translate the Ancients in France

Many of these processes are actually reinforced by the way that translation of the Ancients is taught and learned in France. To be clear from the beginning, the French learning system will not be compared to others, nor shall any prescriptive statements be made on how the way we teach should be changed, the following is simply an elaboration based on personal experience. It is founded to a large extent in my individual learning progression, and the way French institutions work, or at least the way they worked when I was learning Ancient Greek and Latin11.

Let us first examine the two main translation exercises used to teach and evaluate the progress of students in ancient languages (and, more generally, in most if not all languages in the French university system): the « version » and the « thème12 ». Both are very straightforward translation exercises: translating into French and from French, respectively. However, the skills developed in both these exercises will be used in different ways by the student after their curriculum, as in the domain of ancient language translation, one almost always translates from an ancient language into a living one (version), and almost never the other way around (thème)13. Moreover, during my time as a student, and in most curricula in the classical domain, there is no translation courses other than the version and the thème. It is therefore from these sole academic experiences, far from professional expectations, that a translator of ancient languages enters their profession. That is why it is necessary here to underline the importance of these two exercises. Both exercises will nevertheless be practiced by almost all Classics students, for two main reasons:

- both exercises are seen as windows looking into ancient languages, and as ways, in and of themselves, to learn these languages. The objective is not only to learn how to translate, but also to learn how to read the Ancients in their original tongue. This is, in the absence of any living native speakers, our only available means of doing so: we may only practice the language through the cultural production of the dead.

- both version and thème exercises are parts of the agrégation de lettres classiques, the very competitive examination for the selection of Classics teachers in France, and something of a rite of passage for everyone aiming to become a scholar in the domain. Not all students of the Classics will necessarily pass the agrégation, but all will hear about it, often as early as in their first year: since the agrégation is one of the most difficult assessments that a student can come up against, teaching targets it, on the grounds that thus preparing for the worst will allow them to do their best in all other circumstances, or in simpler situations. After all, most of those who will not attempt the agrégation will still pass the CAPES (another, less selective, teacher selection examination), in which, although thème is no longer evaluated, version is still present.

Thus translation of ancient languages in France is structured by these two important competitive tests. Most French translation handbooks specialize in either the version or the thème; they also generally make explicit references to the CAPES, the agrégation, or both, in their titles14. In his introduction, Guy Lacaze (LACAZE, 1999, 7-20) constantly addresses his readers as “candidates” (and not as “students”). This leads to a number of consequences:

- from day one, students prepare to translate a supposedly unknown text, in a limited time frame (four hours for the agrégation), with the objective, given the competitive framework, not simply of producing a translation, but of producing a better translation than any of those produced by other candidates. To paraphrase Yves Mounier, my first Ancient Greek teacher, such competition is not won, it is survived: confronted for very little time to a complex text, the objective is not so much to to make a perfect translation, as it is to not lose as many points on your final mark as the person next to you. This means that candidates will use specific translation methods that are not necessarily the most suited to longer translations: they will focus on speed, since omission of a passage, or not finishing a translation during the CAPES or agrégation almost always means failure; or they will concentrate their efforts on certain types of more “expensive” errors that will have more influence on their final mark, etc. This is the reason for many of the translation errors made by agrégation laureats: one could go so far as to say that it is essential to not be quick in order to do quality long-term translation, since the objective is not to be better than anyone any single occasion, but to dedicated oneself to overall quality, which is arguably more difficult;

- the members of these examination juries need to agree on the criteria of that constitutes good or bad practice in translation. This amplifies the use of well-known and normalized translation structures, including the norms that make a text old-sounding: the agrégation is not considered to be the proper occasion for innovation and linguistic boldness. It demands, and with good reason, rigor and deference towards ancient texts. The extent of this normalization is such that later translations, in professional context, are affected, since most of them are produced in a scholarly context, where precision is paramount. The point of a doctoral thesis or a new volume of the Guillaume Budé collection15 is not (only) leisure reading: it is to produce a tool for students, teachers, and other scholars. Since the agrégation entails mandatory reading and an Ancient Literature examination, and part of the program changes every year, new translations in the Budé collection need to conform to its standard, further discouraging innovation (which, of course, is not always necessary, and may only very rarely even present an interest: the point here is simply to give the contours of the situation), and amplifying the deference bias that we saw in part 1 of this paper. In some of the oldest volumes of the collection, these pressures may have actually helped to discourage simple literal translations.

The most extreme example of the euphemization in this context is the translation from Latin into French of the first verse of the 16th poem of Catullus. The translation analyzed here (LAFAY, 1932) has been revised by Simone Viarre in 2002 : commenting Lafay’s work is, however, interesting, insofar as many public libraries still only possess the 1932 edition, or make it available next to its 2002 counterpart. Thus the first contact with a translation of Catullus is often a contact with the older edition. Here is some context: Catullus is a Roman poet of the beginning of the IInd century CE, whose work is centered on love (both romantic and physical) and the mourning of his brother. He is a poet whose aesthetics are full of contrasts, ranging from deeply moving to the very crass . We must ask the reader to continue reading with a warning, as we will now delve into an extremely vivid description of sexual acts. Catullus begins poem number 16 in Lafay's edition, by the following: “Paedicabo ego uos, et irrumabo...,” which was translated by Lafay (LAFAY, 1932, 13) into “[Je vous donnerai des preuves de ma virilité]”, or “I will give you proof of my manliness”, between brackets, indicating a non-literal translation. This is understandable, as this verse is particularly rude, and a more faithful translation could be: “I will sodomize you and stick it in your mouth”. It is difficult faced with such a case to translate with the respect due to the Ancients: Catullus, in his time, was not necessarily aiming for respectability anyway, or at least not in the way we think of it today. How much does the deference we feel towards the Ancients (that we saw in part 1), reinforced by the normalization of translation induced by the CAPES and the agrégation, exert its influence on translators? I have encountered this fascinating epistemological question in discussions between classicists, but rarely asked or addressed in a scholarly context.

3. Concrete translation problems concerning books 75-77 of Dio's Roman History

Let us now focus more precisely on my current doctoral work. Some of the problems mentioned above do not really affect it that much. Dio has adopted a deliberately classical, Attic Greek style (FREYBURGER, 1997, 26) which makes his work perfectly adapted to the reverence bias that we tend to adopt: he mostly uses a serious, respectful tone16. However, the later books, e.g. books 75 to 77, that I am currently working on, pose specific problems due to the fact that they only survive in fragmentary or summarized form. In other terms, we do not have access to a direct copy of the said books: we can only work from the traces of them left in other works which used them in one way or another.

We can thus find traces of Dio's work in the fragments of Petrus Patricius' History, compiled in the Excerpta de Sententiis by order of the Byzantine Emperor Constantinus VII Porphygenetus. Petrus has reworked some of Dio's phrasing, though, as we can see in the fragment number 132 (BOISSEVAIN 1906, 260) that pertains to book 76 of the Boissevain edition (BOISSEVAIN, 1901, 353-354) :

Ὅτι Πλαυτιανὸς οὕτως ἐπ' ἀδείας διῴκει τὰ πράγματα ὥστε ἐκεῖνον δοκεῖν εἶναι αὐτοκράτορα, τὸν δὲ αὐτοκράτορα ἔπαρχον· ὥστε ἐκεῖνο εἶπεν πρὸς τὴν βουλὴν Σεουῆρος ὅτι « ὁ βασιλεὺς εὔχεται πρὸ αὐτοῦ ἀποθανεῖν »…

My translation follows, with the words I have put in bold characters simply transcribed:

Plautianus managed public affairs with such impunity that he seemed to be the autocratôr, and the autocratôr seemed to be the prefect, to the point that Severus said this to the Senate, that « the basileus wants to die before [Plautianus]… »

I could not translate, at first, the words autocratôr and basileus, because both refer to the same person (Severus, whose authority is weakened by his leniency towards Plautianus), and, also, have been known to translate the same latin word, « imperator ». In actual fact, we know that autocratôr is the word that Dio uses for it (FREYBURGER, 1997, 129-131), and that it had been widely used in this meaning before him (MASON, 1974, 119). However, basileus was also used as a translation of the imperial title, as early as the beginning of the IInd century CE in inscriptions (MASON, 1974, 120). We thus know that both terms refer to the same person and the same title; we also know that basileus must be a word used by Petrus rather than Dio (FREYBURGER, 1997, 132). However, since we only have fragments to work from, we still should translate Petrus' words in the form in which we have obtained them. What to do? Do we use synonyms even though both of these words translate the exact same Latin title? Do we write « emperor » for both words, making the difference between them in the Greek original disappear? I chose to use synonyms, translating autocratôr by « emperor », since I know this is the proper term in Dionian context, and used a vaguer, less precise term (« souverain », i.e. « sovereign » or « monarch ») for basileus. This choice can, though, be debated on methodological grounds: it is a current, open problem.


There are thus a few key differences between translating ancient languages as a student, and translating them as an aspiring scholar:

- as a learner, translation, by ourselves or others, is the only way to have access to ancient texts, and these translations have some particularities due to the status of ancient languages as opposed to modern ones (more particularly, an apparently inevitable bias of deference);

- the way we learn, centered around short format, canonical texts, the study of ancient literature in critical editions, and the preparation of the CAPES and the agrégation, is not necessarily adapted to all the aspects of longer exercises in translation, and further normalizes the form of our translations, sometimes to the detriment of the stylistic variety of the original texts;

- there are practical problems that we face further into our careers (texts transmitted only through fragmentary or indirect sources, for example) that we are not taught how to overcome; we thus tend to solve them by tacit imitation of our predecessors, sometimes without much methodological questioning.

Some of these questions are being addressed currently, but further, broader studies are also called for in the field of ancient literature translation. More specifically, it seems necessary that scholars of Classics dedicate some of their efforts to the larger field of translation studies, so that general methodology is not taken for granted once we have passed our agrégations.

Note de fin

1 Cassius Dio was a Greek historian and Roman senator from Bithynia. He wrote an eighty-volume Roman History at the beginning of the IIIrd century CE.

2 2We do not have to time to detail here the process of establishing an ancient text, we can however point the reader to an excellent article by Jean-Baptiste Camps (CAMPS, 2015), a specialist of the medieval period, which does a good job of explaining the history and controversies in the domain of scholarly text editions.

3 3I am also aware of Pierre Laurens' article for the Bulletin de l'association Guillaume Budé, 2019, vol. 1, titled « Dignité de la traduction », which I have unfortunately not had the occasion to read.

4 4I do not take into account either the administrative or the liturgic uses of Latin and Ancient Greek in Catholic or Orthodox Christianity, nor the Latin academic writing of the Bibliotheca Teubneriana as they are not relevant to literary translation, which is my primary focus.

6 6Literally « unfaithful beauty », i.e. a translation that aims to please the reader by adapting the general content of the original work rather than translating it exactly or literally.

7 7Steiner does acknowledge that linguistic and cultural differences (or proximities) make the translator's task more difficult (STEINER, 1998, 489 ; 491 ; 492-500), but he does not provide a systematic solution to this problem, nor an analysis of how these relationships change with time.

8 8See the posterity of the myth of Medea, for example : after the play of Euripides written in 431 BCE, Corneille (in 1635) and Jean Anouilh (in 1953), among many others, gave their own version of the tragedy.

9 9Compare the translation of Lysistrata in the Collection des Universités de France and the one done by Victor-Henry Debidour who chose to use, at times, the argotic language of his time, for a clear example.

10 10Work of ancient authors are now edited in Greece with a translation into modern Greek (SKOUTEROPOULOS, 2002) whereas Shakespeare is still played in its original language. Translations into modern English do exist, but they are harder to find than the original work.

11 11I will be referring to these institutions by their French names, when there is no readily available equivalent in English.

12 12The Robert & Collins bilingual dictionary doesn't provide exact equivalent to these terms : for « thème », it just gives « translation » (ROBERT & COLLINS, 2020, 986). The case of « version » is a little more complicated : the dictionary (ROBERT & COLLINS, 2020, 1045) gives the expression « unseen (translation) » in its French-to-English section. However, if we search for « unseen » in the English-to-French part of the dictionary, we find (ROBERT & COLLINS, 2020, 2178) that it translates into « version sans préparation ». Thus the term « unseen » does not seem to apply to the fact that the student is translating into their native language, but rather to the fact that they have not studied this text beforehand with their teacher.

13 13There are some Latin or Greek translations of popular works, generally aimed at students or curious readers. Some of the Harry Potter series has been thus translated (the reader will find the references of the first volume in the bibliography). One can also find, mainly secondhand or in some high school libraries, a Latin version of the comic book series Asterix for which I haven’t been able to find coherent publication dates. These books being rather rare, and obviously not aimed at native readers, my point stands.

14 14E.g. for the version exercise: Manuel de version grecque: à l’usage des classes de concours ENS Fontenay/Cloud, Ulm, CAPES et agrégation internes et externes (LACAZE, 1999) and, for the thème exercise: Le thème grec de la licence à l’agrégation (LEBEAU, 2008).

15 15The Collection des Universités de France, also called « Collection Budé » by classicists, is a collection presenting ancient texts with an introduction presenting the history of their transmission through history, the text in its original language, and a translation. It is extensively used for the study of ancient languages and authors.

16 16Dio does have a humourous side, which he develops more in the later books of his Roman History, when he tells anecdotes as a direct witness of the event. See for example Roman History, 77, 8 (BOISSEVAIN, 1901, 363).

Citer cet article

Référence électronique

Ludovic Thérond-Debat, « Translating the later books of Cassius Dio's Roman History into French: a few methodological considerations », La main de Thôt [En ligne], 10 | 2022, mis en ligne le 09 mars 2023, consulté le 24 juillet 2024. URL :


Ludovic Thérond-Debat

Université Toulouse II Jean Jaurès – PLH-CRATA

Articles du même auteur