Voicing 'The Simpsons' from English into French: a story of variable success
Nigel Armstrong, University of Leeds
In this article I compare the use of spoken sociolinguistic and other linguistic variables in the English and French versions of some episodes of the animated cartoon show 'The Simpsons' from the point of view of their deployment, in English and French, to represent social-regional differentiation in the speech of some of the characters in the show, as well as differentiation in character based on voice features. I explain firstly how social identity is mediated through social-regional accents in UK and US English compared to the French of France, examining subsequently how linguistic features carry across from English to French from the point of view of translation 'loss'. I then consider some examples of the use of voice quality in the show, again looking at how successful its exploitation is rendered in translation.
Socio-linguistics, dubbing, translation loss, voice quality
1. The Simpsons
The animated cartoon show is something of a phenomenon, enjoying both popular and critical esteem. The most florid period of its popularity is now past (it first appeared in its current form in 1989) but it continues to attract large audiences on terrestrial and satellite TV channels in countries speaking English and French, the two languages of interest here. The more serious devotees analyse and catalogue various aspects of the show on websites and discussion lists. Perhaps the most obvious reason for the show's appeal to a wide audience, as well as to a more or less obsessed group of cognoscenti, is the fact of the show's functioning on two levels: the straightforward knockabout appeal of the animated cartoon, designed principally for children, and the attraction for adults of erudite references, irony, pastiche, self-reference and other devices.
2. Variable language in English and French
The use of variable language in 'The Simpsons' provides a further level of sophistication in addition to those mentioned above. Linguistic variation, whether in pronunciation, grammar or vocabulary, differentiates speaker groups categorised by sociological criteria such as age, sex or gender, social class, region and ethnicity. The fact of variation is of course axiomatic in sociolinguistics, but in the discipline the aims of studying variation are various; as various, indeed, as the programmes of individual researchers. For example, the Labovian 'variationist' method pioneered by Labov (Labov 1966, 1972), by analysing variable language data using quantitative methods, attempts to formulate the general principles of variation and change that govern how linguistic change penetrates linguistic contexts and spreads socially. Many of the assumptions of Labovian sociolinguistics are based on results drawn from US and UK English, Latin American Spanish and Canadian French. Sociolinguistic variation in the French of France has been rather little studied compared to these languages. A further research programme in sociolinguistics is therefore comparative, with the aim of establishing whether the patterns of variation that have been reported in the French of France conform to those reported in other languages (Armstrong 2001).
In the present hypothesis, we are essentially comparing, in a translation perspective, some examples of voices in 'The Simpsons' that refer to the UK and US spoken dialect patterns with what obtains in France, to see what the fit looks like. In the UK especially it is hard to separate out regional origin from social origin, however this second attribute is defined: social class, level of education, social trajectory, social ambition are all interlocking factors that influence how UK English speakers regulate the regional components of their accents, whether in the short or long term. Put another way, there is a correlation between regional origin and (say) social class such that regional origin is increasingly detectable in speech as one goes down the social-class continuum, and vice versa - the higher the speaker's social class (level of education, etc.), the more attenuated the regional accent, to the extent that the most prestigious accent, 'Received Pronunciation' (RP), is sometimes characterised as 'regionless'. Recent changes in the direction of levelling seem to be flattening this hierarchy. From the perceptual viewpoint, empirical evidence shows that a English listener can identify one of the well-known urban accents very quickly: Harms (1961) reporting the results of an early test that sought to elicit from a panel of listeners judgments of American English speakers' social status on the basis of their speech alone, reported that most of the listeners required only 10-15 seconds to estimate the social provenance of fellow-locutors' accents. There seems to be no reason to suppose that UK listeners require more time. In the UK, the best known accents are those of London, Newcastle, Liverpool, Manchester, Birmingham, the loose Yorkshire conurbation (Leeds, Bradford, Sheffield). Less prominent accents are probably identified with reference to the broader regions: north, midlands, south, etc.
The point of immediate comparison here is of course the US dialect pattern, although we shall see that UK English accents are also exploited in 'The Simpsons'. For our present purposes we can state briefly at this point that alongside a generalised, socially relatively neutral US accent, there are distributed among others the rather low-prestige southern accent, as well as the prestigious north-eastern New England accent. We discuss relevant aspects of the US dialect pattern in more detail below, when we look at the speech of individual characters in the show.
Another way of describing sociolinguistic patterning is by saying that speakers signal their identity, or localisation in what Hudson (1996: 207) has called a 'multi-dimensional [social] space', by situating themselves, on the one hand in relation to social values (perhaps principally level of education) associated with the supra-regional standard language variety available to them; and on the other, to the localised vernaculars associated with solidarity-based, 'home-team' values. A further important dimension, the counterpart to speech production, is evaluation, which is integral to the sociolinguistic nexus that interrelates the extra-linguistic factors (social class, age, sex, ethnicity) influencing variable language. Quite obviously, a linguistic act of identity does not take place in a vacuum, but is presented to a hearer. To cite Eckert (1989: 248): 'Labov's original (1966) findings in New York City clearly lined up socio-economic class, style, sound change, prestige, and evaluation on a single axis'. One aspect of the evaluative dimension is that at a first encounter, a listener attempts to identify a speaker's social characteristics by analysing (among other features) their language, as Trudgill's (1995: 1-2) example is designed to illustrate: two strangers in a train compartment will use accent among other clues to place their vis-à-vis socially. A further step is one of evaluation: listeners form a normative judgment on the basis of their identification. The dimensions of class (as well as sex and age), style, sound change, prestige, and evaluation are indissociably linked: for example, prestigious language varieties are more conservative, in the sense of representing an earlier state of the language, and are used by speakers who broadly are older and more highly placed socially.
We need furthermore to distinguish these 'acts of identity' as they take place on the three linguistic levels of pronunciation, grammar and vocabulary. It is pertinent to mention here Hudson's suggestion (1996: 43-45) that speakers express different aspects of their social identity on different linguistic levels, such that that while morpho-syntactic (grammatical) variation tends to be suppressed across languages, phonological variation is cultivated so that speakers can express various acts of identity. As Hudson expresses it (p. 43): 'it could be that we use pronunciation in order to identify our origins', while 'we may use morphology, syntax [and vocabulary] in order to identify our current status in society, such as the amount of education we have had.' This argument is summarised and further developed 'very tentatively' as follows (p. 45):
syntax is the marker of cohesion in society, with individuals trying to eliminate alternatives in syntax from their language. [...] Pronunciation reflects the permanent social group with which the speaker identifies. This results in a tendency for individuals to suppress alternatives, but in contrast to the tendency with syntax, different groups suppress different alternatives in order to distinguish themselves from each other [...].
Hudson's phrase 'in contrast to the tendency with syntax' is in line with his suggestion that variation in grammar tends have a different kind of social significance from pronunciation, with the result that the suppression of all grammatical variants is aimed at; in contrast with pronunciation and vocabulary, where non-standard alternants are kept alive.
Is this true of France? We need obviously to distinguish firstly between northern and southern France. As is well known, there is a considerable difference between the broad accent groups of these two large regions. Our present focus of interest is on 'northern' or 'standard French', or le français de référence - there is no satisfactory compact term. We can call the variety of French of interest here français d'oïl: the non-southern French of France, spoken in the area broadly north of the Garonne and the Massif Central, excluding extreme areas such as Alsace in the east and the Breton-speaking west of Brittany, where influence from non-Romance languages is at work. A further important qualification is that we are concentrating on urban varieties: so, 'urban oïl French'. The term oïl stands in contrast to oc (as in langue d'oc); the terms derive from the Old French words for 'yes' and symbolise the broad two-way dialect division that characterises the French situation.
There is some evidence to suggest that the pronunciation of urban oïl French is quite highly 'levelled' compared to the UK situation (Armstrong 2001, chapter 2). By levelling we mean the reduction of pronunciation or other variables that distinguish speaker groups: young-old, male-female, middle-class-working-class, speakers from different cites. Pursuing still the UK-France comparison, we need also to make a distinction between 'qualitative' and 'quantitative' pronunciation variables. An example of the former type from UK English is the /a/ vowel in the second syllable of 'decade'. In one of the broadest varieties of Tyneside (north-eastern) English this can be pronounced [e:], where the colon shows lengthening. In Cockney the quality is very different, more like [ai]. These variation phenomena are often phonetically arbitrary, in the sense of being resistant to ease-of-articulation explanations. In contrast, 'quantitative' pronunciation variables differ in terms of presence or absence, rather than in differences of vocalic or consonantal quality. Quantitative pronunciation variables are sometimes referred to as 'deletion phenomena'. The French sequence y a pas de doute is a reduction of il n'y a pas de doute and could be rendered in English as 'There's no doubt', also of course an elided form and probably having a social-stylistic value that matches the French form reasonably closely.
In the UK and US, qualitative variables generally have a clearly recognised regional-social distribution and are often negatively perceived, at least from the viewpoint of the 'ideology of the standard' (Milroy and Milroy 1991), the attitude that sees the standard as the only real language and all other varieties as imperfect approximations to it. By contrast, deletion phenomena are shared in varying degrees by all speakers and hence have limited socio-stylistic value, since speaker-hearers seem to be tolerant of the need to economise time and effort so long as meaning is not compromised. Facility of translation of a stretch of non-standard pronunciation depends on whether phonetically arbitrary variables or deletion phenomena are in question; since deletion phenomena are found across languages, then approximately equivalent translation effects are achievable. We shall see below that qualitative or arbitrary pronunciation variables represent a formidable obstacle to the translator.
Corresponding to the relatively levelled nature of urban oïl French pronunciation, it appears that social-stylistic variation in French has been to some extent displaced to the grammatical and lexical levels. It has been pointed out by several scholars (Hudson 1996: 45; Chambers 1995: 51-52) that grammatical variation is less likely to be quantitative than variation in the sound system (for reasons discussed in the following paragraph), but one can question the view that variation on this linguistic level by its nature virtually always shows polarised patterns of variation; this view may reflect a strong bias towards English in the existing data. Certainly, the findings available (e.g. Wolfram 1969; Cheshire 1982) suggest that grammatical variation in English is usually polarised, with the variable use of some grammatical features present in working-class speech, but almost totally avoided by middle-class speakers.
A rather dissimilar situation in French is suggested by the rather fragmentary variationist data available, represented for instance by findings reported in Valdman (1982) and Coveney (1996). These findings suggest a sociolinguistic distribution of certain French grammatical variables that resemble phonological variables in their non-polarised patterning. Thus it may be that a language such as French, whose phonology has been successfully levelled, may not conform to the tendency to suppress grammatical variation; this tendency is suggested by Hudson (1996: 44-8) as a general one across languages. As we shall see below, the relative levelling of French pronunciation puts difficulties in the way of the oral translation of some social-regional accents that are used with rather subtle effect in 'The Simpsons'.
Regarding lexical variation, an important difference between the two languages is the large number of non-standard terms available to French speakers. Pairs like bosser and travailler, bagnole and voiture, rigolo and amusant show that lexical variation includes nouns, verbs and adjectives. A comparison between English and French highlights the fact that although there are of course plenty of slang terms in English, the extent of the phenomenon seems wider in French, both in the number of casual or informal terms used and the number of people who use them. Very many French speakers will refer to their car as their bagnole when talking to family and friends. There appears to be no equivalent term in English that is so widespread socially. Similarly, if we consider non-standard synonyms in English for 'eat', we see that although the commonly used French word bouffer is of course a more casual term than manger, it seems to have exactly the same denotational reference. Instead of 'eat' an English speaker might say 'scoff', but this seems to add something as well as casualness - 'scoff' implies greed of the eater as well as informality of the speaker. The set of lexical pairs is fairly large; one researcher (Armstrong 2001, chapter 7) counted 237 pairs of this kind in a corpus of spoken French. From the translator's perspective, the difficulty resides in the very fact of the ready availability of the pairs of French lexical alternants, such that much French slang is perhaps less marked or visible than in English.
This is perhaps the counterpart in lexis of the relative lack of social variation in French pronunciation and grammar. The following example, taken from "The Last Temptation of Homer" (Episode 1F07, first broadcast 12 September 1993), seems to illustrate this. To punish Bart Simpson for a piece of bad behaviour, Bart's class teacher Mrs Krabappel announces that for the rest of the term Bart will be called on to answer all questions in class. She then asks Bart to pronounce and define the term 'photosynthesis', and it turns out that he cannot read the word written on the blackboard. Astonished, Mrs Krabappel exclaims:
Mrs Krabappel: Is it possible that all your misbehaviour and miserable grades have been caused by a simple vision disorder?
Bart: You mean it ain't me noggin, it's me peepers? Oh well, that's just loverely!
This is conveyed in the TT as follows:
Mrs Krabappel: Est-il possible que ta mauvaise conduite et tes résultats médiocres soient dus à un simple problème de vue ?
Bart: Ça viendrait pas d ma caboche mais d mes mirettes ? Ça alors, c'est chouette !
At a less complex level we see Bart transposing his teacher's standard utterance into a more demotic register, to achieve a fairly straightforward effect of comedy through bathos. One feature of interest here is the relative closeness of equivalence of ST and TT in the standard speech given to Mrs Krabappel. By contrast, the speech given to Bart in the two versions differs considerably; Bart's ST speech is voiced in a Cockney accent, as the spelling 'loverely' is designed to indicate. A further Cockney feature is 'ain't', and while the rather old-fashioned slang terms 'noggin' (head) and 'peepers' (eyes) have no particular regional localisation, their old-fashioned flavour combines with the Cockney features to evoke (in the present writer's mind) references to 'Oliver Twist' and 'My Fair Lady'. The effect is to add complexity to the ST through the use of marked social-regional language. At first sight this looks like sheer whimsy, of a kind that is not uncommon in the show, but one can suggest that the attribution to Bart of a stretch of archaic Cockney is meant to reinforce our perception of the 'Artful Dodger' side of his character. This is done by the use of stereotypical language features - ones that are no longer in current use but continue to be associated in speaker-hearers' minds with a clearly defined social or regional category of speaker. To be fair to the translators of 'The Simpsons', the use of a social-regional stereotype here defies the translator's art, since the UK dialect pattern, where region and social class are indissociably intertwined, has no close parallel in France. Their best attempt is to match the bathetic effect of the original by putting the teacher's standard French alongside some slang, which, as discussed above, has no particular localisation in time or space, is indeed the property of virtually all French speakers on account of the readier availability and copiousness of slang in that language (Armstrong and Hogg 2001). It certainly seems plausible that caboche and mirettes are less marked, because probably more frequent in French, than 'noggin' and 'peepers' are in English.
3. Variable language in 'The Simpsons': two individual studies
Any attempt to analyse the employment of variable language use in 'The Simpsons' needs to take into account the fact that the show conveys representations of certain aspects of English-speaking culture, this latter term taken in its broad anthropological sense. Representations, in this sense of the term, are by their nature factitious, and cannot be thought as conveying a veridical portrait of the behaviour of (for example) a given stratum of society. This might be thought self-evident where an animated cartoon is in question, but 'The Simpsons' is sometimes referred to as an 'animated sitcom', a definition implying a focus in the show upon human issues that can be expected to involve the exploitation of variable language, as well of course as other types of behaviour. The show seems to be the first of its kind to offer a relatively complex human-interest element along with the humorous and fantastic effects that are inseparable from the animated cartoon, if only because the medium is capable of exploiting the non-veridical. Concerning the human-interest dimensions dealt with in the show, classic social polarities between male and female, young and old, middle class and working class and different ethnicities, are all featured in 'The Simpsons', in a way that is more sophisticated than what is found in obvious predecessors such as 'The Flintstones'. At the same time, these polarities are treated in 'The Simpsons' in a way that is proper to what might one simply call 'art;, the approach that often exploits the purely creative and ludic use of language and other socio-cultural practices. For instance, the eponymous family is portrayed as belonging to a relatively under-financed stratum of the working or lower-middle class, but social-class differences are rarely approached in a didactic or 'committed' way. Rather, they are the source of an often sophisticated form of humorous entertainment, which indeed frequently finds expression in the use of linguistic variables.
The foregoing sketch is intended to give some idea of the linguistic complexity that characterises 'The Simpsons'. On another level, sociolinguistic studies of non-spontaneous speech of the type found in the show need to recognise that the speech, as well as being scripted (collaboratively, in the case of 'The Simpsons') is produced by voicing artists, some of whom voice more than one character. For instance, Dan Castellanata voices such disparate characters as Homer Simpson, Mayor Quimby and Groundskeeper Willie. Despite the astonishing versatility and virtuosity of the voicing artists in their mimicry of social-regional accents, it remains true that some accents are produced in a stereotypical way, by exaggerating certain of their most salient features.
Bearing in mind these caveats, in what follows we discuss some examples of the use of variable language from English and French versions of 'The Simpsons', looking at variable phonology and voice quality. We do this by studying the language given to some of the principal characters in the show.
3.1 Charles Montgomery Burns
'Monty' Burns occupies a prominent place in the Simpsons universe. He is the local magnate, owner of the nuclear power plant where Homer works, and is portrayed as being 104 years old. He is almost always represented negatively, as a heartless, grasping megalomaniac. His speech reflects his great age: it is peppered with uncommon terms like 'crapulence', as well as old-fashioned exclamations like 'fiddlesticks!' and 'huzza!'.
Aside from old-fashioned lexis, a more complex effect of characterisation is achieved by attributing to Mr Burns a pronunciation that shares characteristics of a UK English accent and the anglicised, upper-class US New England accent. This is in a long US tradition that gives movie villains an English or anglicised accent, reflecting a negative stereotype that continues presumably to be widespread in the US. The strategy in the French version is to give Mr Burns an upper-class accent sometimes referred to as 'seizième', referring to the prosperous sixteenth district of Paris. Mr Burns in the French version of 'The Simpsons' is therefore highly placed socially, but perhaps integrated more closely in being given an accent that forms part of the French dialect pattern. This is in contrast to the original version where his anglicised accent demarcates him more sharply from the rest of the Simpsons universe, by attributing to him an accent that is marginal in the US dialect pattern.
3.2 Mayor Quimby
'Diamond' Joe Quimby, the mayor of 'Springfield', the generic US small town where most of the action of the show is set, is given the most highly focused accent in the show. He has a Boston accent, and along with its distinctive intonation, the accent is designed to refer to the speech of John F. Kennedy and perhaps the male members of the Kennedy 'clan' more generally. Indeed, a very explicit allusion is made in one episode where Mayor Quimby asserts, in a speech to his constituency, 'Ich bin ein Springfielder'. It seems likely that the accent is employed to represent Mayor Quimby as belonging archetypically to the political class. He is portrayed as a populist, corrupt womaniser, reflecting perhaps the revisionist view of JFK. Honey (1989: 136) characterises Kennedy's accent as 'the Boston hyperlect, the American equivalent of the poshest form of British (marked) RP'. Honey points out that Kennedy, following advice from his political team, attenuated his accent in order to gain wider acceptance among voters, showing how a highly marked upper-class accent is capable of arousing hostility in an egalitarian age.
The most distinctive feature of the Boston accent is that, unlike other prestige US accents, it is 'non-rhotic', or to use further jargon, it has no 'post-vocalic /r/'. So whereas in the standard US accent, pronunciation of /r/ in the sequence 'fourth floor' (relevant Rs underlined) communicates what in the sociolinguistic terminology is called 'overt prestige', the Boston accent has no /r/ in these phonetic contexts; or more precisely, it is generally absent, given the variable nature of pronunciation features of this kind. Post-vocalic /r/ is a good illustration of the fact that speakers can employ linguistically entirely arbitrary phonological items in a way that is highly charged socially: thus rhoticity or presence of /r/ is used by overtly prestigious speaker groups in the US (the Boston area excepted), while almost exactly the opposite pattern obtains in the UK. In UK English the vowels in the sequence 'fourth floor' are long monophthongs, maintaining the same the same vocalic quality throughout the syllable. In Quimby's Boston accent, at least as conveyed in the show, the vowel in 'floor' is perhaps exaggeratedly realised as a triphthong, where the vowel changes quality twice within a syllable. This pronunciation might be rendered in spelling impressionistically as follows: 'flow-euh'.
Consultation of the intuition of native speakers indicates that the French strategy in portraying Quimby is to attribute to him the accent of a small-time crook, an unsuccessful gangster of the type featured in films like those by Tarantino. This is achieved not through the use of an accent, but through a deep and harsh, 'gravelly' voice quality. This is clearly negative stereotyping, and to that extent is broadly equivalent to the effect aimed at in the original version, but the interesting difference is that while the original version exploits an accent that has quite clear (if perhaps erudite) reference to a veridical accent, the French version has recourse to a stereotype that derives to some extent from a cinematic convention used to define a stock character in a certain genre of film. There must inevitably be an element of social reality behind the cinematic stereotype, since stereotypes do not arise from the void, but the connection between reality and art is more tenuous in the French version. Clearly, a close match to the Quimby voice in French is impossible, since the way in which the accent pattern in France maps on to social organisation is so different from the US and UK. It seems surprising nevertheless that the French editorial team concerned with overall strategy did not at least attempt a like-for-like match; an obvious parallel is the accent or rather idiolect of Valéry Giscard d'Estaing, the former French president who is well known for his peculiarity of often pronouncing an /s/ as a 'sh' sound. The transposition of 'Quimby as Kennedy' to 'Quimby as Giscard' might have represented a closer cultural match in view of Giscard's rather tainted reputation in France; it is certainly true that matching one political figure with another seems more satisfactory, however approximative, than substitution of 'cheap crook' for 'tainted politician', the chosen French strategy. It may that the French team failed to decode the social significance of the Quimby accent; it is far from transparent to many English telespectators. A further consideration is that Giscard's reputation is currently more or less rehabilitated.
We mentioned earlier that social-class differences in 'The Simpsons' are not treated in a politically committed way. It is noticeable nevertheless that the two characters in the show having very clear positions of financial advantage and political power are demarcated from most of the other characters in the show through the attribution of a social-regional accent (as opposed to through voice quality, which we discuss below).
4. The use of voice quality: two individual studies
The notion of voice quality, although an intuitive one for non-linguists, is nevertheless in need of careful definition if it is to be useful. Crystal (1991: 376) defines voice quality as 'the permanently present, background, person-identifying feature of speech'. Crystal points out that this 'feature' can be defined linguistically; in terms of combinations of characteristics like tempo, loudness, pitch, nasality, whisper, lip-rounding, breathiness, etc., but also in impressionistic, affective terms such as 'cheery', 'haughty', 'sullen', etc. (Crystal's examples). We can remark that a third, quasi-linguistic terminology is also possible, through the use of terms like 'gravelly', 'strangled', 'plummy', 'throaty', etc. We use the term voice quality in the former, 'permanently present, person-identifying' sense in what follows, although we will be obliged also to use impressionistic terms of the third kind just mentioned.
4.1 Marge Simpson
Marge, the long-suffering wife and mother of the dysfunctional family, is portrayed in both the English and French versions with a hoarse voice quality, to use an impressionistic term. This is 'permanently present' and 'person-identifying', and is perhaps designed to convey the extent to which Marge is tired and harassed by her lynchpin role in staving off the disasters that frequently threaten the family. Hoarseness is a voice quality that transfers successfully across the two languages of interest here, as it conveys characteristics of much the same sort in English and French. This is perhaps because hoarseness in this particular context has a very concrete, physical base in the reality of Marge's situation: we can imagine her raising her voice a good deal to make it heard above the ambient noise produced by children and inadequate husband. Such physicality transcends cultures (or at least cultures that tolerate a raised voice in such contexts) and hence translates directly.
4.2 Homer Simpson
The principal character is usually conveyed in a negative light: selfish, overly sentimental, greedy, ignorant, slow on the uptake. As in much artistic enterprise, this portrayal is adhered to only when it suits the authors' purposes: Homer can show fair intelligence when a joke depends on it, as in the exchange between him and his clever daughter Lisa:
Lisa: Dad, do you know what a rhetorical question is?
Homer: Do I know what a rhetorical question is!
This exemplifies the literary paradox of the independence of the text over its characters, a phenomenon that is at least as old as 'Macbeth', where one of the murderers is given a line of poetic description of the twilight. Most of the time, Homer is however portrayed as being of rather limited intelligence, and his voice quality is an important component in conveying this. It is of interest that voicing artists have rather little insight into the articulatory detail of their portrayal of the characters they voice. In a tenth-anniversary programme shown on British television, Dan Castellanata, the artist who voices Homer, remarked that initially he selected a 'Walter Matthau' articulation which he described a being 'adenoidal or whatever' - 'nasal', in the jargon. Castellanata went on to recount that he found this voice quality expended too much energy, was therefore difficult to sustain, and that it conveyed too limited a range of emotion. In Castellanata's word's the voice kept dropping 'down into the throat' and was preferable because in that way it could express both positive and negative emotion, and also because it sounded 'dopier'. In articulatory terms, Homer's voice quality often seems slightly 'pharyngeal'; the 'dopey' effect is achieved by constricting the pharynx, the part of the throat cavity behind the back of the tongue. Pharyngeal voice tends to convey a comic effect in English, although it is a permanent setting in some languages, the Scandinavian for example, which may to some extent explain the slightly comic perception that some nations have of speakers of Danish, Swedish and Norwegian.
A further feature that contributes to the portrayal of Homer as not being among the most gifted is his often rather slow speech rate. Speech rates, considered as a permanent individual characteristic, can vary considerably across individuals broadly irrespective of other factors such as formality of speech situation or the stress of emotion. A fast articulation rate appears to be positively regarded in many societies, as indicating competence in general and perhaps a high level of intelligence in particular (Giles 1992: 133). It is no doubt significant that rustic speech is stereotypically portrayed as occurring at a slow articulation rate. Yet another feature is an exaggerated intonation pattern, which gives a childish effect. These features combine to undermine our capacity to take Homer seriously. All transfer successfully to French.
5. Concluding remarks
Since we have already discussed the implications of our findings in some detail above, we confine ourselves here to summarising them briefly. We have concentrated here on two aspects of oral translation. We examined firstly two examples of language that are so saturated in the culture to which they refer as to rule out literal translation, if indeed we can admit the concept of anything like literal translation where a spoken accent is in question, other than the standard perhaps. These examples are the accents of Monty Burns and Joe Quimby, and we have seen that the French strategies for conveying these accents vary in the closeness of cultural transposition they achieve. The French version of Mr Burns's accent is a qualified success, in that the 'upper-classness' of the character is conveyed, but in a rough-and-ready way that attributes to him superiority but not otherness, or only to a lesser degree. The transposition of Quimby's accent falls rather shorter of what is achieved in the case of Burns; Quimby is portrayed in the French version in cruder and more attenuated terms. The general conclusion concerning the oral translation of accent from English into French is fairly obvious, though worth reiterating: the local is related to the social in US and UK English in a more intimate way, one that has been dissociated in the variety of French in question here. The translators therefore lack the resources to achieve close equivalence, although it is of interest that the standard accent of Mr Burns transfers fairly successfully; this is because elevated or standard speech more or less by definition is unmarked as to region. We wish to avoid here normative judgments on the French translation strategies; these must be examined in the socio-cultural context in which they take place.
Voice quality, by contrast, transfers more closely from English to French. One is tempted to say that this is because its physicality lends it universal properties, and this is perhaps true in the case of Marge Simpson. A wider sample of translations would need to be examined to check this (the programme is broadcast in about 100 countries). This is also true of the rather more complex portrayal of Homer through voice quality. We stated above that a voice quality having temporary or idiolectal effect in one speech community can be a permanent setting in another. This is the case with pharyngeal voice, and so it seems likely that its successful transfer from English to French is fortuitous, one of the strong cards in the translators' hand.
Armstrong, Nigel (2001). Social and stylistic variation in spoken French: a comparative approach. Amsterdam and Philadelphia: John Benjamins.
Armstrong, Nigel and C. Hogg (2001). "The Pollyanna principle in French: a study of variable lexis." Journal of Pragmatics 33 (11): 1757-85.
Chambers, Jack K. (1995). Sociolinguistic theory. Oxford: Blackwell.
Chambers, Jack K., and P. Trudgill (1998). Dialectology (second edition). Cambridge: Cambridge University Press.
Cheshire, Jenny (1982). Variation in an English dialect. Cambridge: Cambridge University Press.
Coveney, Aidan B. (1996). Variability in spoken French. A sociolinguistic study of interrogation and negation. Exeter: Elm Bank.
Crystal, David (1991). A dictionary of linguistics and phonetics (third edition). Oxford: Blackwell.
Eckert, Penelope (1989). The whole woman: sex and gender differences in variation, Language Variation and Change 1: 245-67.
Giles, Howard (1992). "Attitudes to language'", in The Oxford International Encyclopedia of Linguistics volume 4, Bright, W. (ed.), 132-34. Oxford: Oxford University Press
Harms, L. S. (1961). "Listener judgments of status cues in speech." Quarterly Journal of Speech 47: 164-68.
Honey, John (1989). Does accent matter? The Pygmalion factor. London: Faber and Faber.
Hudson, Richard A. (1996). Sociolinguistics. (second edition). Cambridge: Cambridge University Press.
Labov, William (1966). The Social Stratification of English in New York City. Washington, DC: Center for Applied Linguistics.
Labov, William (1972). Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press.
Milroy, James and Lesley Milroy (1991). Authority in Language (second edition) London: Routledge.
Trudgill, Peter (1995). Sociolinguistics (fifth edition). Harmondsworth: Penguin.
Valdman, A. (1982). "Français standard et français populaire: sociolectes ou fictions?" French Review 56 (2): 218-27.
Wolfram, Walt (1969). A Sociolinguistic Description of Detroit Negro Speech. Washington: Center for Applied Linguistics.
Nigel Armstrong is Senior Lecturer at the University of Leeds and researches into sociolinguistic variation in the contemporary spoken French of France. He is the author of numerous publications including Social and stylistic variation in spoken French: a comparative approach, Amsterdam and Philadelphia: John Benjamins, 2001.
He can be reached at email@example.com