Dialects and nonstandard languages are undoubtedly translators’ pet peeves – or mountains to climb for the most determined. All of them agree that translating linguistic variation is an achievement in itself, if not simply impossible. Any translator that decides to get to grips with such texts should first consider his task from a sociocultural and historical perspective before focusing on the linguistic aspects proper. Theorists are very much aware of the inequalities between languages, between those considered dominant and those of minorities that are more often than not passed over, and Lawrence Venuti’s works have laid bare the ethnocentric trends in literary translation (VENUTI, 1995).

The linguistic variety that the present study is devoted to is the variety of English spoken in Ireland, often called Hiberno-English (the origin of the prefix is the Latin word Hibernia, which means Ireland). This variety is defined by Dolan as follows: “In Ireland, Hiberno-English means that you have two languages in a kind of unruly shotgun marriage together, fighting all the time over the centuries, for syntax, pronunciation, vocabulary, idiom” (AMADOR-MORENO, 2007, 214). Novelist Roddy Doyle, whose works are teeming with some of the most realistic illustrations of Hiberno-English, has become one the most famous authors in Irish literature: “Like Synge with his Aran Islanders or O’Casey with his slum-dwelling Dubliners, Doyle has the gift of transcribing precisely (and enjoyably) a vernacular dialect.” (WHITE, 2001, 9). It is mostly by capturing his characters’ voices as faithfully as possible that Doyle has been able to grant literary recognition to a whole geographical and social community. Since his characters’ speech defines them as part of an easily recognizable world, I propose to examine the French translations of several extracts from five works by Roddy Doyle: The Commitments (1988), The Commitments (translated by Isabelle D. Philippe, 1996), The Snapper (1990), The Snapper (translated by Bernard Cohen, 1996), The Van (1991), The Van (translated by Isabelle Py Balibar, 1996), The Woman Who Walked into Doors (1996), La femme qui se cognait dans les portes (translated by Isabelle D. Philippe, 1997), Paula Spencer (2007) et Paula Spencer (translated by Isabelle D. Philippe, 2012)1. The choice of this corpus seems rather relevant since, while it offers a homogeneous sample of the author’s writing – a trilogy and two volumes focusing on the same character, it has been translated into French by three different translators. The analysis of the translations is therefore meant to study the various strategies worked out by the three translators and bring to light their similarities and differences so as to assess each one’s outcome.

1. Standardisation

Standardisation has often prevailed in the translation of dialects. Alexandra Assis Rosa calls it “normalization”, which she defines as “a corresponding change from source text stigmatized or less prestigious literary varieties to the most prestigious variety in the target text: the standard” (ASSIS ROSA, 2012, 87). This strategy is the most commonly used in the literary translation of linguistic varieties, where the need to abide by the rules of written literary usage originates both from automatic responses deeply rooted in the practice of translation itself and from the impact of the publishing process – a dimension that has to be considered in the present study. The translator is therefore bound to take into account the part played by the source literature in the target culture. The more in minority the language is – and hence unknown outside of the source culture , the more normalised its translation will tend to be. Massimiliano Morini talks about his own experience as translator of Sunset Song by Lewis Gibbon, pointing out how cultural background influenced his work (MORINI, 2006, 123-140). In non-English-speaking countries, Anglophone literature is often seen as a unified whole and little attention is paid to the specific origin of the language, whereas this aspect is of utmost importance in terms of identity issues. In Italy, for example, the Scottish specificities of the writing are disregarded as a minor characteristic of a text which is mainly identified as belonging to the English-speaking area, as Morini reports:

Though Italian readers are generally aware that Walter Scott is a Scottish writer, they will not think of Scottish literature as a separate tradition from English […] literature, but merely as English literature written in Scotland. (MORINI, 2006, 128)

Likewise, Ritva Leppihalme considers the value of the source text within the target literary community, admitting that the works of an author who is not very well known (or recognised) tend to go through a systematic standardising process in translation (LEPPIHALME, 2000, 247-269). She emphasizes the importance of the target readers’ expectations, pointing out that the more linguistically “marked” the language is, the more reduced the readership tends to be:

From a distance, the linguistic details that make up the richness of a particular author’s language can be nearly invisible. A translator may want to highlight them only if all other circumstances make the effort worth his or her while, and in that case, the cost will include a reduction in the size of readership if we assume (as, realistically, I think we must) that the majority of general readers would not prioritize linguistic individuality. (LEPPIHALME, 2000, 267)

This notion is in keeping with the place usually assigned to the source text in literary hierarchy: a work supposed to be “mainstream” will be more easily adapted to the target context so as to reach a larger number of readers; accordingly, its linguistic peculiarities will be the first to be rubbed out to make way for a standard, more accessible version of the language. In the present case, it can be alleged that the publishers did not intend to reach an elitist readership made of Irish culture specialists. Rather, they aimed to take advantage of both the film adaptations’ commercial success and the 1993 Man Booker Prize winner’s fame to sell the novels in France. This can be confirmed by the choice made by the publishers at Robert Laffont to have The Barrytown Trilogy translated by three different translators, giving priority to deadlines rather than stylistic consistency. It would then appear that importance was placed on production speed, which suggests that the novel was meant to be sold quickly to the largest number of people, who would have access to the trilogy’s three volumes at the same time. It would seem that the “mainstream” dimension prevailed here, while the novels’ Irish origin was somewhat overshadowed by the more general label of Anglophone literature, paving the way for standardisation.

1.1 Grammar and syntax

A survey of the translations shows that the original text’s grammatical peculiarities are lost in the French versions, hence the disappearance of the marked forms. This gives birth to an average syntactic arrangement that shows no difference from the norm.

No literal translation of “incomplete perfectives2” (HICKEY, 2005, 44) is possible since it would fail to be understood in the target language. The only option available is to normalise the syntactic construction to obtain more natural expressions in standard French:

TS	D’yeh expect me to cough up every time the man thinks Leslie done somethin’? (p. 96)	Tu voudrais que j’allonge du fric à chaque fois qu’il « pense » que Leslie a fait quoi que ce soit ? (p. 259)
TV	I seen yeh, said Jimmy Sr, again. (p. 111)	Je vous ai vues, point. (p. 286)
PS	I seen you drinking Coke. (p. 99)	Je t’ai vue boire du Coca. (p. 189)

The French language is known for its lack of flexibility, which does not allow for unacknowledged grammatical structures. This is confirmed by examples of inverted indirect interrogative forms that are typical of Hiberno-English3. Their translation was systematically standardised in the target text to obtain a syntactically acceptable sentence:

TS	She remembered Jackie was asking her was she alright. (p. 29)	Elle se rappelait que Jackie lui avait demandé si elle allait bien. (p. 69)
TWW	I asked Jack was she in his class. (p. 82)	J’ai demandé à Jack si elle était dans sa classe. (p. 146)
PS	She wonders have they remembered. (p. 15)	Elle se demande s’ils y ont pensé. (p. 24)

The analysis of the above examples highlights the various obstacles raised by the distortions of grammar and syntax in nonstandard language when one attempts to translate it into a notoriously nonflexible language. Hiberno-English possesses a variety of forms for the second-person plural subject pronoun “you” (HICKEY, 2005, 270-271), used by Doyle as “ye”, “yis” and “youse”4. The translator is faced with the challenge of translating each of them when the only pronoun available in French is “you”. It seems impossible to invent a new pronoun, hence the necessity of standardisation, as proved by the examples below in which the pronoun has been almost systematically5 replaced by “vous”:

TC	—Wha’ tracks are yis doin’? Jimmy asked. (p. 13)	- Et sur quel morceau vous travaillez ? demanda Jimmy. (p. 15)
	—Fuck up, youse, said Jimmy. —Tha’ was years ago. We were all fuckin’ eejits then. (p. 52)	- Vos gueules, vous deux, trancha Jimmy. C’était il y a des années. On était tous des petits cons à l’époque. (p. 123)
TS	—D’yis know wha’ they had me doin’ today, do yis? Yis won’t believe this. (p. 63)	- Vous savez ce qu’ils m’ont fait faire aujourd’hui ? Je vous le donne en mille. Vous ne devinerez jamais. (p. 162)
	—Fuck off, youse, said Bimbo. (p. 41)	- Oh, allez vous faire foutre ! dit Bimbo. (p. 99)
TV	—I’ll have to leave yis now, I’m afraid, he told them. — We’re a bit understaffed in the kitchen. (p. 134)	- Il faut que je vous quitte maintenant. On manque un peu de personnel aux cuisines. (p. 347)
	—Will youse go with Billy, lads? he asked Muggah McCarthy and Pat Conlon. (p. 19)	- Vous voulez pas aller avec Billy, les gars ? demanda-t-il à Muggah McCarthy et à Pat Conlon. (p. 39)
TWW	—Now Father O’Hanlon has a few words he wants to say to yis. (p. 112)	- A présent, le père O’Hanlon veut vous dire quelques mots. (p. 205)
PS	What’s it like for them? Are yis not freezing? (p. 39)	Comment vivent-elles ça ? Vous ne gelez pas ? (p. 70)

It is obvious that this lack of dynamic equivalence is due to a grammatical void in the French language where only “vous” is used as a second-person plural pronoun. Besides, as Françoise Wuilmart points out, the invention of grammatical forms is not encouraged in French-language literature: “in French you have to be a poet or receive the blessing of the Academy to attempt any linguistic change” (WUILMART, 2007, 133).

1.2 Visual dialect

Visual dialect offers a phonetic transcription of the linguistic variety employed by characters. What makes nonstandard language so difficult to translate is that it swerves from a “norm” that is not only lexical and grammatical but also orthographical. Linguistic variety may therefore be graphically marked out, particularly when accents are recreated in literary works. This process, known as “visual dialect” (KRAPP, 1925), allows the author to have his characters speak in such a way that the reader can hear a variety that already appears at the syntactic level. Visual dialect, which “designates any form of deviating from the orthographical conventions of a given language, aimed at representing a transcription of spoken language” (LOOCK, 2012, 41), is part and parcel of the written construct of linguistic variety. Although it makes the literary production of speech more faithful, visual dialect is nonetheless a paramount obstacle when it comes to translation. Since this quasi-phonetic transcription of language rests on pronunciation, its rendering in another language is highly problematic. The difficulty increases when the target language is French since this language offers fewer possibilities than other, more flexible linguistic systems: visual dialect “represents a real challenge for the translator, especially when he has to translate into French, a language known as normative and far less liable than English to stray from the norm” (LOOCK, 2012, 40). The English language offers a variety of options thanks to its orthographic flexibility:

Few languages display such a tenuous relationship between sound and orthographic representation as there is in English, hence, the use of eye dialect sensu stricto may not be feasible in the target language. (BRETT, 2009, 50)

The characters’ target-language speech in the present corpus is devoid of any specific markers of their ways of speaking, with the exception of one example – occurring twice in the translation of TS – that shows an attempt at translating the different ways of greeting each other into French:

—Hiyis, she said when she got there. —Oh, howyeh, Sharon. —Hiyeh, Sharon. —Howyeh, Sharon. —Hiyis, said Sharon. (TS p. 12)	- Sa-lut, dit-elle en arrivant à leur portée. - Oh, saluuuut, Sharon. - Saaaalut, Sharon. - Sa-lut, répéta Sharon. (p. 23)
—Hiyis, she said. —Hiyeh, Sharon. —Ah howyeh, Sharon. —Hiyis, said Sharon. (TS p.33)	- Sa-lut, dit-elle. - Saaaalut, Sharon. - Oh, saluuuut, Sharon. - Sa-lut, répéta Sharon. (p. 77)

These two examples show that the transcription does not concern the same semantic unit. The translation of the English phrase « how are you? » is just the French equivalent of “hello”. It would have been possible to use a small range of questions, such as, “Comment ça va ? Ça va ? Tu vas bien ?”, but the translator decided to give priority to the spoken language in order to not set aside the peculiarities of the source text. While offering no faithful rendering of any specific accent, the repeated vowels do point at some difference from the norm. Though no echo of any recognisable French accent can be found in the target text, the reader will certainly wonder about the occurrence of specific phonetic markers. However, the analysis of other visual dialect markers in our corpus proves that this option could not be chosen systematically. The spelling “oul’”, which corresponds to the Hiberno-English pronunciation of “old” (DOLAN, 2006, 179), disappeared completely since the French adjective “vieux” does not offer any phonetic alternatives:

TC	He’s a fuckin’ oul’ fella. (p. 37)	Mais merde, c’est un vieux ! (p. 89)
TS	They’d be better company than your oul’ fella annyway, wha’. (p. 11)	En tout cas, ce sera une meilleure compagnie que ton vieux père, non ? (p. 22)
TV	It’s cos they’re afraid their oul’ ones’ll catch them if they drink in the Hikers, Anto told Bimbo and Jimmy Sr. (p. 17)	C’est parce qu’ils ont peur de se faire piquer par leurs vieux s’ils vont au Hikers, cafta Anto. (p. 41)
TWW	I sound like an oul’ one. (p. 17)	Je parle comme une vieille. (p. 23)
PS	They’re young. She’s the oul’ one again. (p. 39)	Elles sont jeunes, Paula est encore la vieille. (p. 70)

The language of Doyle’s speakers is characterized by its phonetic peculiarities. Their accent can be heard throughout the text, as shown by the apocope of word-final consonants as in “wha’”, “tha’”, “righ’” or  “-in” endings. However, the language that can be found in the target texts has been orthographically normalised:

TC	Is tha’ abou’ righ’, Joey? he asked. (p. 39)	Ça va à peu près, Joey ? demanda-t-il. (p. 94)
TS	—Is tha’ wha’ yis’re callin’ him? —That’s righ’, said Jimmy Sr. (p. 20)	- C’est comme ça que vous l’appelez ? - Oui, confirma Jimmy Sr. (p. 43)
TV	Say we go into town, righ’; we go into town an’ we try an’ get into one o’ those disco bars, righ’? (p. 25)	Si par exemple on allait en ville, hein ? On va en ville et on essaye d’entrer dans un de ces fameux bars disco, okay ? (p. 59)
TWW	Fuck off, you, righ’. (p. 36)	Lâche-moi, toi, d’accord ? (p. 53)

The strategy used by Doyle to stress the characters’ accents does not stop there. In the whole corpus the spelling « yeh » is used instead of “you”, and the pronunciation typical of the Dublin area is materialised by the double consonant that transforms “any” into “anny”, altering the vowel sound in the first syllable of such words as “anyway” (“annyway”). The novels are also scattered with occurrences of nonstandard syncopated spellings: “oney” (“only” TC), “prob’ly” (“probably” TS, TV), “def’ny” (“definitely” TS, TV), among others. The disappearance of phonological markers, quite usual in translation, is often debated in translation studies. Our own observations and often unsuccessful attempts bring us to the conclusion that it was impossible to meet the requirements of the source text.

1.3 Results

Antoine Berman’s observation that “translation is a powerful centralizing anti-dialectal agent” (ASSIS ROSA, 2012, 88) is backed up by the examples quoted above. Alexandra Assia Rosa sees language as a number of concentric circles where each circle moves further away from standard language – from spoken language to geolect, then to sociolect. When the grammatical and phonetic markers of Hiberno-English as it is spoken in Dublin are omitted, the target text offers a neutral form that tends to bring the language back to the centre of the linguistic sphere, where specificities are the fewest. Our own survey tends to confirm this hypothesis since, although not having consulted each other, the three translators spontaneously resorted to similar strategies to get over the obstacle raised by nonstandard language. The danger of this homogeneity is the possible erasure of the particular and the strange. The Irish geolect, its accents, lexicon and grammar disappear from the target text to become standard French, devoid of its specific geographic flavour, which is highly detrimental to Doyle’s text where, as Lisa McGonigle points out, “the connection between voice and identity is of paramount importance in this text” (MCGONIGLE, 2005, 169). The limits of the translation of dialects or nonstandard languages are quite obvious, as shown by a number of examples, and many specialists have brought to light the losses suffered by the source text in this process. While Gregory Rabassa, translator of Gabriel Garcia Marquez, declares that local or regional idioms are impossible to translate (FAWCETT, 1997, 121), Kaen Bruneaud-Wheal points out that "traditionally, when mentioned in theoretical approaches to translation, nonstandard language is considered by translators and academics as verging on the untranslatable" (BRUNEAUD-WHEAL, 2010, 5). Even those translators who have seriously considered this issue during their work – which is not necessarily the case for any of them – have regretfully noticed that, in spite of their efforts, the outcome does not meet their expectations: "normalization and simplification seem once again to be the unconscious toll to be paid if a text is to be transferred from one language and culture to another" (MORINI, 2006, 135). Besides, the French language, which is codified by the Académie Française, does not readily make for distortion, as Karen Bruneaud-Weal points out, affirming that the English language "lends itself more easily than French to creation, hence to distortion" (BRUNEAU-WHEAL, 2010, 13), to which Annie Brisset adds that the translation of sociolects in French is hampered by the existence of a "linguistic void in the normative system of its literature" (BRISSET, 1990, 299). The obvious danger of normalisation is the possible disappearance of some essential element of the narrative, or even some new delineation of the characters: "The effect of such linguistic normalisation is to distort the social identity of the characters of Doyle’s text" (HORTON, 1998, 428).

If the rearrangement of the syntax aimed to fit the target language runs counter to the nonstandard language which is grammatically completely rubbed out, it is nonetheless possible to preserve the characters' spoken language. This strategy does away with the specificities of the dialect without normalising it completely. As Karen Bruneaud-Wheal remarks: "This partial standardisation fails to represent speech as "other" and brings it back to a local stereotype (a somewhat familiar language), according to an ethnocentric process" (BRUNEAU-WHEAL, 2010, 12).

2. Partial standardisation

The aim is to create a new linguistic area, an in-between zone, so that the foreign items may be brought into the target language and be easily understood. This system brings to light the gap between the nonstandard language and the standards of the target language while not depending on any acknowledged linguistic variety. In most cases, when an equivalent dialect cannot be found in the target language, the translator tends to rely on the synonyms that exist in the various language registers where the colloquial is often seen as a possibility to translate the nonstandard. The method called "vertical translation" is commonly used by translators to make up for the "horizontal" linguistic variety, which means that geographical variation is rendered by register variation. The character's voice is then altered, and its regional and social features are lost, while bits of coarse language are added (though not present in the original). Here the context and the character's features are both lost. Loock takes the example of the French translation of Winston Groom's Forrest Gump, where "the overuse of slang [...] eventually turns the character's speech into a sometimes vulgar one" (LOOCK, 2012, 53).

Many specialists in translation studies think that, if form itself is not given priority in the case of nonstandard languages, then the specific colouring that it gives to the speech must be, as David Horton points out:

It is widely agreed that (generally speaking) it is not the linguistic form, but the sociolinguistic function of dialectal forms which should be observed (and, ideally, preserved) in the act of translation. (HORTON, 1998, 418)

What has to be taken into account is not the purely linguistic peculiarities, but the various effects the specific speech marking achieves on the source reader , and the way they can be reconstructed for the target reader. This reflective approach means that the translator must not only be familiar with the source cultural context, but also make sure to grasp the author's purpose. As a matter of fact, language betrays a geographical and cultural context as well as social and generational origins; the more marked it is, the more specific the "clues" are. In the present corpus, the nonstandard language points to geographical data (Dublin) as well as sociological ones (the working class) and the transcription of a colloquial language variety which seems to have prevailed in each of the three translators' approaches.

2.1. Orality

The translators sought to enhance the specific use of "sure" that occurs at the beginning of some sentences6. The examples listed below reflect the translators' tendency to make use of "literary orality" in their rendition of colloquial speech. The characters' speech is intended to be realistic, very close to literal transcription, using some phrases that are typical of oral speech.

TC	—I s’pose yeh are, said Outspan. —Fuckin’ sure I am. (p. 14)	- Non, t’as pas tort, concéda Outspan. - Et comment que j’ai pas tort ! (p. 18)
TS	—These yokes aren’t as nice as they used to be, said Jimmy Sr. —Sure they’re not? (p. 31)	- Ces machins, ils ne sont plus aussi bons qu’avant, dit-il. Hein, j’ai raison ? (p. 73)
TV	Ah sure, said Bimbo. —That’s wha’ they’re supposed to do at her age. She’s lovely. (p. 56)	- Ah ben, sûr. C’est de son âge. Elle est adorable. (p. 140)
PS	—We can’t give them Coke, she tells Paula, after she’s handed over the bottles. —Sure we can’t, Popey? (p. 72)	- On ne peut pas leur donner du Coca, informe-t-elle Paula, après avoir distribué les bouteilles. Hein, qu’on peut pas, Popeye ? (p. 134)

Besides, the target text confronts the reader with the kind of realistic oral speech that characterises Doyle's style in the whole novel; the French translation resorts to such common features as truncated negation where the adverb "ne" is omitted:

TC	- Tha’ sounds like me arse, said Outspan. – But I’m sure you’re righ’. (p. 10)	- T’y vas un peu fort, déclara Outspan. Mais t’as pas complètement tort. (p. 16)
TS	- You’re stainin’ the carpet. (p. 36)	- T’as fait plein de taches sur le tapis. (p. 86)
TV	- Did yeh stay in a hotel? Jimmy Sr asked him. (p. 27)	- T’as logé à l’hôtel ? lui demanda Jimmy Sr. (p. 65)
TWW	- You’re not to eat any. (p. 107)	T’as pas le droit d’en manger. (p. 195)
PS	- Where were you? (p. 42)	Mais où t’étais ? (p. 76)

Accordingly, the vowel in the subject pronoun "tu" was elided and replaced with an apostrophe occurring before a vowel:

TC	- Tha’ sounds like me arse, said Outspan. – But I’m sure you’re righ’. (p. 10)	- T’y vas un peu fort, déclara Outspan. Mais t’as pas complètement tort. (p. 16)
TS	- You’re stainin’ the carpet. (p. 36)	- T’as fait plein de taches sur le tapis. (p. 86)
TV	- Did yeh stay in a hotel? Jimmy Sr asked him. (p. 27)	- T’as logé à l’hôtel ? lui demanda Jimmy Sr. (p. 65)
TWW	- You’re not to eat any. (p. 107)	T’as pas le droit d’en manger. (p. 195)
PS	- Where were you? (p. 42)	Mais où t’étais ? (p. 76)

Some truncated forms such as "d'acc" or "d'ac" (instead of "d'accord"), or else "déc'" ("déconner") do sometimes occur, and the relative pronoun "que" may also be omitted:

TC	—Righ’, said Jimmy. —Are yis ready, girls? (p. 34)	- D’acc, dit Jimmy. Vous êtes prêtes, les filles ? (p. 69)
	—What’s it to you if she was? said Bernie. (p. 39)	- Qu’est-ce ça peut te faire ? Riposta Bernie. (p. 88)
TS	—Meetin’ the lads, yeh know. See yeh, righ’. (p. 75)	- Je dois retrouver les potes, tu comprends. A plus tard, d’ac ? (p. 199)
TV	—An’ you can sleep with Darren. How’s tha’? (p. 33)	R’marque, tu pourrais dormir avec Darren, qu’est-ce t’en dis ? (p. 81)
PS	- For fuck sake, Ma. Was he good-looking? (p. 43)	Sans déc’, maman. Il était beau ? (p. 78)

A close study of the approach to orality shows that it is more elaborate in TV than in the other works of the corpus. Indeed, while all the translations rely on the methods expounded above, Isabelle Py Babibar's translation makes extensive use of contracted forms that go beyond the elision of the vowel in "tu". The "e" in the first-person plural disappears, replaced with an apostrophe, and the contracted spelling "r'garde" is used as an equivalent of the omitted preposition in "look it"7 (DOLAN, 2006, 153). The target text features a huge number of such contracted phrases as “ j’viens”, “ j’vais”, “j’sais”, “j’veux”, “j’dis”, “j’te” (and so many others of the same kind), along with the truncated form that results from the omission of the subject pronoun in sentences beginning with "z'allez" ("vous allez") or "y a" ("il y a"). The extracts below are examples of how this device is employed throughout the novel:

—Gettin’ locked tonigh’, men? said Anto. (TV p. 19)	- Z’allez être bouclés, ce soir, les gars ? (p. 39)
—There’s loads o’ things he can do, said Jimmy Sr. (TV p. 47)	- Y a des tas de choses qu’il peut faire. (p. 116)
—Home, said Jimmy Sr. —I’m knackered. (TV p. 53)	- J’vais me pieuter. J’suis crevé. (p. 133)
—Oh, here we go, said Jimmy Sr. —Look it. (TV p. 66)	- Oh, r’garde un peu qui vient. (p. 167)
—But there’s no fuckin’ water, Jimmy Sr said again. (TV p. 84)	- Mais, bordel, puisque j’te dis qu’y a pas d’eau ! (p. 212)

The use of a certain degree of colloquial language, conveyed by phonetic transcription, was given prominence by the three translators (especially Isabelle Py Balibar) in order to emphasize the importance of oral features. The result of this strategy is a hybrid language, neither standard nor dialectal. The characters' voices present no specific characteristics and are therefore deprived of any individual identity. As Rohan Anthony Lewis points out, the translator does not draw his inspiration from any specific language existing in the target culture but "from the lectal range of one specific community" (LEWIS, 2003, 419). This is not a process of complete normalisation since, as seen above, some of the markers are still present, but the loss of the Irish flavour cannot be denied. This strategy, aimed at increasing the degree of colloquial language to make up for the loss of geographical identity, is very likely to be used when one translates into French since the variety of French registers offers many possibilities.

2.2. Registers

The use of register to compensate for the lack of appropriate dialect in the target culture is very common in translation. As Karen Bruneau-Wheal observes, nonstandard language is rendered in French by "markers and turns of phrase that are typical of spoken language, especially of a colloquial type" (BRUNEAUD-WHEAL, 2013, 94). Similar conclusions are drawn by Marion Beaujard in her study of the French translation of Roddy Doyle's novel A Star Called Henry: "the translator has to deal with the specificities of the language by replacing the regional features with register markers" (BEAUJARD, 2013, 138). Doyle's novel raises an issue at once when it comes to the translation of registers since the original is scattered with colloquial language or slang; besides, its linguistic variational markers display a whole range of sexual or scatological terms that are typical of Dublin English, as Hickey indicates (all the terms quoted below occur in our corpus):

There can be little doubt that the reputation of vernacular Dubliners for being generally foul-mouthed rests on the large range of words having to do with various intimate bodily functions. […] boner ‘erection’[…] (gob)shite ‘person disapproved of’ […] gooter ‘male organ’ […] jacks ‘toilet’, langer ‘male organ’ (HICKEY, 2005, 141)

While it seems obvious that the target-language register should be as colloquial or slangy as the original, the translator does not necessarily have to resort to coarse, or even aggressive, expressions to make the language more colloquial. Although such words as "connard" or "enfoiré" can indeed be used to translate "bollox", these words sound much too derogatory for "eejit"8 (DOLAN, 2006, 89). Strangely enough, the French word "idiot" has been left aside and replaced with a variety of rude words such as "con", "taré" in TC (one occurrence of "idiot" out of six possibilities), "dingue", "rat" "triple nœud" in TS (no occurrence of "idiot"), "cinglé", "enfoiré", "enculé" in TV (no occurrence of "idiot"), "tarée", "demeurée" in TWW (three occurrences of "idiot"), or else "naze", "débile" in PS (no occurrence of "idiot"). It is possible, therefore, to conclude that, in the translator’s eyes, using "idiot" to translate "eejit" would have meant erasing something from the original. However, the loss is not a semantic one – "eejit" is the Hiberno-English pronunciation of "idiot" in standard English – but a dialectal one, and one fails to understand why dialectal differences should be rendered by register differences. We may wonder whether the target reader is likely to understand that "cinglé" or "taré" belong to some marginal, nonstandard language. The same question arises for the word "garda"9 (DOLAN, 2006, 110) which occurs both in its Gaelic plural form "gardaí" and with its anglicised spelling "guard", a non-colloquial word which has been borrowed from the Gaelic and simply refers to "policier" or "police" in their general sense. And yet, this word has often been rendered by "flic" in TS, TV and PS, which raises the question whether linguistic variation can be signalled by the use of colloquial register. As Jean-Michel Déprats observes, Fouad El Etr has opted for a more substandard language and slangy words to translate linguistic variation in one of his translations of Synge's works, a device which proved detrimental to the original style:

Synge's language is not made of colloquial or slangy words which do not reflect the peculiarity of his idiom. Indeed, although the expressions that occur were – and sometimes still are– common in Ireland, they remain strange or unusual in the standard language that is English. (DEPRATS, 1998, 72)

Similarly, any French reader will immediately grasp the colloquial dimension of the word "flic" but will have no indication of its variation from the linguistic norm. It seems better to resort to the strategy advocated by Isabelle D. Philippe, who chose to preserve the original term both in PS ("les Guards", p. 144) and in TWW ("un Guard") where she adds the following note: "Nom des policiers irlandais" (p. 11).

It seems quite obvious, however, that slang was woven into the translation of “messing”10 and “messer”11 (SHARE, 2005, 236). The three translators opted for a colloquial or coarse register to render these terms which have nothing to do with slang in the original language. Let us examine in detail what this approach entails in the target text. The examples drawn from TC are the transcription of the conversation between Jimmy Junior and his friends:

TC	Declan Cuffe stared across at Jimmy while he sent his cigarette to the side of his mouth. —You startin’ somethin’? he said. […] —Ah, cop on, said Jimmy. —I was only messin’. (p. 14)	Declan Cuffe dévisagea Jimmy en déplaçant sa cigarette vers le coin de sa bouche. - Tu me cherches ? […] - Ah, arrête, dit Jimmy. Je déconnais, c’est tout ! (p. 25)
	Mickah laid down the rules as each of them passed the table into the hall. —Anny messin’ an’ I’ll kill yeh, righ’. (p. 50)	Mickah énonçait le règlement au moment où chacun d’eux passait devant la table pour entrer dans la sale. - A la moindre emmerde, je vous tue, compris ? (p. 127)

The use of colloquial words has no real incidence on the target text since the young people’s language often tends to be quite coarse. This contrasts with the examples drawn from TS, where George Burgess’s and Darren’s words undergo drastic changes:

TS	I’m sorry, Jimmy. Really now. On the Bible. I was just messin’ with the lads, yeh know. (p. 55)	Je suis désolé, Jimmy. Non, vraiment. Sur la Bible. Je déconnais avec les potes, c’est tout. (p. 140)
	He laughed and rubbed his hands and looked around him, laughing. —You’re messin’, said Darren. (p. 97)	Il rit, se frotta les mains, et jeta un coup d’œil à la ronde en continuant à rire. - Tu te fous de moi, accusa Darren. (p. 260)

In the first example, the translation of Burgess’s retort turns the character into some coarse-mannered teenager, whereas no hint of colloquial language occurs in the source text. The second example highlights the danger of this kind of change. Darren is talking to his father who is kidding him, but while in the original version he is just pointing out that he has seen through his little game his reaction in French appears much more aggressive. The use of the impolite expression makes Darren’s retort sound particularly rude and insolent, which alters the interaction between father and son. This alteration is all the more blatant in the examples from TV where the expression used is even more rude:

TV

That would have explained the ketchup she now saw on the ceiling. —Ah no, look —! —Wha’? ⎯⎯Where? ⎯⎯Jaysis, how did tha’ get up there? —I don’t know what you two messers are up to — —We’re not messin’, Veronica, Jimmy Sr assured her. —It’s business. (p. 138)

Ça aurait expliqué qu’elle voyait à présent du ketchup au plafond. - Oh non, regardez ! - Quoi ?... Où ?... Nom de Dieu, comment c’est allé jusque là-haut ? - Je sais pas ce que vous fabriquez, mais comme fouteurs de merde… - On fout pas la merde, Veronica, lui assura Jimmy Sr. On monte un truc. (p. 358)

The text describes Veronica who finds her husband and son learning to juggle with a bottle of ketchup. Although we can easily imagine how surprised and annoyed she is, the slang used in the translation seems exaggerated, especially since in TS and TV Veronica often praises politeness and scolds other family members for their rude manners. Yet in the French version she comes across as just as rude as the others; the maternal image conveyed by the target text is not very flattering and her reaction undoubtedly quite out of line with the mere childish foolishness suggested by the original text. The same process can be observed in the examples from PS quoted below:

PS	—Legend. Babe. She certainly knows her popes. Jesus, Jack, you’re a messer. She wants to hug him. (p. 118)	- Légende. Canon. Elle connaît à donf ses papes. Seigneur, Jack, tu es chiant ! Elle a envie de le serrer dans ses bras.
	She runs. A car honks, twice. She doesn’t look. It’s only some messer. (p. 153)	Elle court. Une auto klaxonne, deux fois. Elle ne regarde pas, ce doit être un chieur. (p. 302)

In the source text, the interaction between Paula and her son that is emphasized in the first example betrays Paula’s connivance and amusement towards her son’s attitude, while the colloquial register employed in translation hints at a reproachful tone, not to mention the rude manners that highly differ from Paula’s behaviour as it is described in the original text. This dramatic change in the protagonist’s character resurfaces in the second example in which Paula’s judgement sounds much more critical in French: the use of “chieur” denotes irritation, while “messer” points at a lack of seriousness. Not only is the characters’ language altered, but the use of a lower register also affects their representation in the target text.

2.3. Result

Although the words and the register used display a range of possibilities that were not grammatically or phonetically available, the outcome is not a faithful rendition of the speakers’ speech or character. This case illustrates why Karen Bruneaud-Wheal deplores the use of colloquial registers as substitutes for regional peculiarities (“while the oral dimension is preserved, its colour is lost”), also denouncing the “tendency to use rude language” (BRUNEAUD-WHEAL, 2013, 106). This “verticalisation” of translation choices is quite suited for sociolects where the use of one kind of register may betray one particular social environment, whereas it turns out to be ineffective for geographic localisation. As Marie Sylvine Muller explains,

If the function of dialect was only the characterisation of the characters’ social background, their words could indeed be rendered by colloquial and popular turns, which would be quite easy to achieve. (MULLER, 1996, 68)

And yet, Doyle’s style denotes both Irish – especially Dublin – origins and colloquial language. This is the reason why this centralising process, although preferrable to the normalisation described before, remains incomplete. It seems essential, therefore, that translators should reflect on the function of identification – cultural, social or geographic – of the dialect.

3. Compensation

It is of utmost importance to take into account the various sociocultural components of the characters’ speech. According to Ake Persson:

When, as a former teacher at a northside school, Doyle makes his characters in the Trilogy speak in a broad northside sociolect, indicated to a large extent through typographical devices, this sends clear sociocultural signals. (PERSSON, 2003, 51)

The point here is not to give prominence to fidelity instead of linguistic form but to reflect on the value of the sociolinguistic clues available in the target language. David Horton declares that “no two dialects can (interlingually) carry the same set of social, ethnological, cultural-stereotypical associations” (HORTON, 1998, 418).

3.1. Slang

The translations present in our corpus display an array of words originating from French slang, defined by Le Larousse as

the set of words used by a living social group, exclusively among group members, as a means of standing out from others or protecting themselves from them (certain trades, higher education schools, prisons or the underworld)12.

The choice of a specific lexical category shows that the translator intends to transfer something from the source-language community into a group that can be identified in the target culture. As Antonella Capra points out, French slang offers a means to set what is expressed apart from standard language:

On a daily basis, the second language used in France, which allows one to swerve from the norm and is employed within a family, a community or a region, is slang, this form of expression with no clearcut boundaries which wholly belongs to official language, but which still possesses a marginal status, as dialects do. (CAPRA, 2014, 4)

The following examples show that the three translators had no hesitation in resorting to slang and even used words hardly known by part of the French population in order to mark off the characters’ speech and bring in a specific social and community dimension.

TC	It’d look deadly on the posters. (p. 9)	Ça cracherait sur les affiches. (p. 13)
	It’s a poxy programme on RTE. A talent show like. (p. 70)	Une émission naze sur RTE. Un genre de télécrochet. (p. 184)

The community identity signalled by slang in the two examples from TC is perfectly suited for the social group that Jimmy Junior and his friends represent. Here we find instances of language as it is spoken by young people who share the same codes.

TS	Does the dog like sandwiches, does he? Jimmy Sr asked her. (p. 22)	Est-ce que ton clebs aime les sandwichs, dis ? (p. 51)
	I’m fuckin’ pissed. (p. 108)	Je suis pétée, putain. (p. 292)

These examples from TS show that slang takes a variety of forms, more or less contemporary, according to the identity of the speaker. Jimmy senior uses a more outdated word (which is the first transcription of “clebs” in the 19th-century Larousse)13, while Sharon, in the second example, uses a much more modern word whose definition in the Larousse14 does not mention the meaning of being very drunk.

TV	Yeh’d ride your missis on it it’s so clean. (p. 82)	Tu pourrais tringler ta bourgeoise dessus tellement c’est propre. (p. 213)
	Your man, look it; don’t let him get past yeh! (p. 106)	Ce taré, fais gaffe ! Le laisse pas te doubler ! (p. 274)

The examples from TV highlight one of the risks of using slang in translation: the language becomes not only colloquial but more aggressive. In the two examples quoted, the French translation of Jimmy Senior’s language turns out coarse (“tringler”) and almost violent towards the driver since the original expression “your man” becomes “taré”. While the recourse to slang is certainly apt for conveying community identity, it may be less efficient in the translation of individual identity.

TWW	I thought he was an absolute ride. (p. 53)	Je le trouvais vraiment canon. (p. 85)
	I was a good fighter myself; I could crease any young one that ever got in my way. (p. 49)	Moi-même, je savais me battre ; j’étais capable d’éclater n’importe quel gamin qui se mettait en travers de ma route. (p. 79)

The examples from TWW illustrate the two tendencies previously mentioned. In the first example, the slangy word “canon” – which is not rude at all but acts as a social marker – is particularly appropriate for the language of a young woman describing a man to whom she is attracted. In the second example, however, slang enhances the rude gesture through the use of the word “éclater” which seems far too violent for a teenage girl.

PS	She was going to sack him last night, but the bollix wasn’t there. (p. 112)	Elle allait le virer la veille, mais le baltringue n’était pas là. (p. 215)
	Same ol’ shite, she says. (p. 105)	La même vieille daube, dit-elle. (p. 202)

The examples from PS show that the target text comes as close as possible to the source text when the slang chosen by the translator reflects what lies at the core of the original. In these two examples, the use of slang signals nonstandard language (i.e., language that swerves from the norm) and the speaker’s specific social background (the expressions used are not necessarily known by all), while remaining just as rude.

To sum up, slang allows the introduction of nonstandard elements into the target language, but it must be selected carefully to avoid making the language too coarse and altering the identity of the characters. However, while it does capture the sociological dimension of the language, it fails to convey the spatial data of the source geolect.

3.2. Back slang

If slang points out a specific social background, it provides only limited geographical information. As a matter of fact, slang in the broad sense of the word is used everywhere in France with no indication of a specific geolect, while the source language and the target language do not often share the same historical or social background. In the case of the present corpus, there is in France a linguistic variety which is proper to the communities that live in the margins of big cities, especially Paris, called verlan (back slang). According to Alena Podhorna-Policka, the creation of back slang is “geographically limited to the suburbs of Ile-de-France (Greater Paris)” and lexical novelties are then “adopted by young people of the suburban areas of other towns as a sign of virtual belonging to ‘street culture’” (PODHORNA-POLICKA, 2006, 47). The translators of our corpus have sometimes resorted to back slang quite spontaneously, as shown by the occurrences of “meuf” (back slang for “femme”) in the examples below. It is interesting to note that the sociocultural dimension of the words prevailed over the meaning: indeed, in TC, “meuf” is the translation of both “mot”15 (DOLAN, 2006, 169) and “brasser”16 (SHARE, 2005, 44), although the second one is much more derogatory.

TC	—The chicks? —Jaysis, Jimmy! —The brassers, yeh know wha’ I mean. (p. 11)	- Pour les gonzesses ? - Bon Dieu, Jimmy ! - Les meufs, tu vois ce que je veux dire. (p. 17)
TC	—All tha’ mushy shite abou’ love an’ fields an’ meetin’ mots […] (p. 11)	- Tous ces trucs à la noix sur l’amour, la campagne et les rendez-vous avec les meufs […] (p. 18)
TV	And five cigars for Jimmy Sr from Aoife, his mot. (p. 41)	Avec cinq cigares pour Jimmy Sr offerts par Aoife, sa meuf. (p. 103)
TWW	He was boasting, of course. Only just married and his mot was already pregnant. (p. 132)	Il paradait, bien sûr. Marié depuis peu, et sa meuf était déjà enceinte. (p. 243)

The presence in the target text of a word belonging to such a distinctive linguistic variety entails a very strong social marking, which inevitably calls to mind the Parisian suburbs, where back slang was born, since “from the seventies onwards, this mode of speech has been successful among teenagers from the housing blocks around Paris” (PODHORNA-POLICKA, 2006, 41). This is in keeping with the geographical and social context of the original works, since the economic similarities between the northern suburbs of Paris and those of Dublin have their roots in the history of the working classes who could not afford to live in town and had to move away to the margins of the two capitals (the Industrial Revolution led the most miserable to move to the north and north-east of the big cities). As Alena Podhorna-Policka explains, back slang has a role to play in “the building of identity” which “enables young people to become integrated, to acknowledge each other and to set themselves apart from other young people and other generations” (PODHORNA-POLICKA, 2006, 42). This probably explains why Isabelle D. Philippe chose such specific words from the Paris area as “ouf” or “kéblo” (both of which originate from back slang, their respective meanings being “fou” and “bloqué”).

PS	—Will I put the kettle on for you? —Yeah; thanks. —This is mad, says Paula. —I can’t cope with it. Pretending. I’m sorry I hit you. I’m sorry. (pp. 104-105)	- J’allume la bouilloire pour toi ? Ouais, merci. - C’est ouf, répète Paula. Je ne peux pas lutter, je fais semblant. Je suis désolée de t’avoir frappée. (pp. 200-201)
PS	—Are all these yours? —No, says Paula. —Some of them. —Have you no good stuff? —Like what? —Stop being thick. The 70s. (p. 133)	- Ils sont tous à toi ? - Non. Quelques-uns seulement. - Tu n’as pas de bonne musique ? - Comme quoi ? - Arrête de jouer la kéblo ! Celle des années soixante-dix. (p. 260)

Since, as Alena Podhorna-Policka explains, “back slang occurs […] as a kind of linguistic signal of the suburbs’ evolution” (PODHORNA-POLICKA, 2006, 42), its use allows one to anchor the characters in a specific geographical and socio-economic background as close as possible to the source culture, although complete equivalence is hard to find: “regional or class-based accents, and all the stereotypes they evoke, are unlikely to have counterparts in other languages” (BRETT, 2009, 50). As a matter of fact, the cultural connotations of back slang have a certain impact on the image of the characters, and the man called Jimmy who calls a woman “meuf” is not quite the same as the man called Jimmy that used to call her “mot”. Although the northern suburbs of Paris and Dublin are economically quite similar, as seen above, their respective evolutions are not the same. Whereas immigration has played a most significant part in the development of French society, this phenomenon was hardly known in Ireland when the Trilogy and TWW were published (only in PS does the new multicultural Irish society begin to emerge). This explains why a number of words relating to immigration in France may sound preposterous in the mouth of Dubliners at the end of the eighties, such as “bled” in TS:

—Yeh can go for a swim with the Pope, said Yvonne. They laughed. —Cos there’ll be fuck all else to do there, Yvonne finished. (TS p. 13)

- Là-bas, tu pourras piquer une tête avec le pape, dit Yvonne. Rires. - Vu qu’il n’y a rien d’autre à foutre dans ce bled, ajouta-t-elle. (p. 24)

Although it refers nowadays to a “usually unattractive village”17, the word “bled” still retains its original North African meaning (“blad” means “country” in North African Arabic) which closely connects it with the history of French colonisation. This example reveals the impact of the introduction of words that are strongly anchored in the target culture. As Brett puts it, “the translator risks transferring the text onto a sociolinguistic plane distant from that of the original” (BRETT, 2009, 51), as the risk of transposing the language into a completely alien context must be examined.

3.3. Conclusion

After pinpointing the devices used by the three translators, our analysis has brought to light a prevailing tendency which consists of centralising the dialect to favour register-marked language. The nonstandard specificities of the original tend to disappear to make way for a regiolect which only occurs through back slang and a linguistic variety which is the transcription of spoken language. Though only partial, the normalising process allows one to tone down the nonstandard dimension of the target text that does not account for the whole range of original constructions. The solution suggested here rests on the creation of a linguistic in-between zone, which David Horton describes as follows: “an option somewhere in between indicating significant deviation from standard norms without attempting to suggest a specific and identifiable language variety” (HORTON, 1998, 418). The strangeness of the language must be enhanced without an easily recognisable form. However, since the translator is not the author himself, he can take very few liberties with the norm, especially in French. According to Loock, “he might be reproached for every specific choice that would deviate from the norm”, since “you have to be an acclaimed author like Balzac or Queneau to indulge in swerving from well-established, rigid norms” (LOOCK, 2012, 55). The translations presented in our corpus do not provide the French-speaking reader with a complete picture of the “Irish language” since the target language may give the impression that what has been translated is standard English. The didactic aspect of translation is, therefore, partly missing since it fails to promote the Hiberno-English dialect, usually unknown by French people who, although well aware of the differences between American English and British English since high school, assume that the Irish speak the same English as the British. The target text might then highly benefit from the presence of a full-fledged peritext that would lay bare the dialectal peculiarities of the source language.