Bernard Shaw's Pygmalion and Alan Jay Lerner's My Fair Lady, as well as the contextual factors which might have influenced the translators' choices. Keywords: ...
How important is the way you say it? A discussion on the translation of linguistic varieties
DRAFT VERSION
Sara Ramos Pinto Faculty of Letters – University of Lisbon University College London Discussions of translation often rely on the concept of meaning -‐ not only the meaning of the words, but also the meaning of the use of certain words in a certain text and context. Moreover, translation always involves a process of identifying the different components of the texts in order to establish a hierarchy of relevance of those elements (see Toury 1980: 38). The priority given to some elements to the detriment of others will have a decisive influence on the choice of certain strategies and the final outcome. The literary use of a dialect in literary texts seems to be a particular good example of that balancing of meaning and prioritization of elements. Not only because of its very localised meaning (both in time and space), but also because it is always embedded in the source text with a communicative and semiotic significance. It can challenge the translator who, when faced with the impossibility of looking for referential equivalencies and formal correspondences, is forced to decide on the importance and meaning of the use of a specific dialect in the text. That decision will define the strategies to be used, which can go from total normalization of the text to a recreation of a linguistic variety in the target text. The purpose of this article is twofold: to present for discussion a model summarising the different strategies acknowledged by different researchers in several case-‐studies; to present and discuss the strategies recognized in a corpus of 12 Portuguese translations of Bernard Shaw’s Pygmalion and Alan Jay Lerner’s My Fair Lady, as well as the contextual factors which might have influenced the translators' choices. Keywords: Linguistic variation, normalization, page translation, stage translation, screen translation
1
1
The flower girl: Ow, eey y∂-‐ooa san, is e? Wal, fewd dan y’ d∂-‐ ooty bawmz a mather should, eed now bettern to spawl a pore gel’s flahrzn then ran any athaht pyin. Will y∂-‐oo py me f’them? (Here, with apologies, this desperate attempt to represent her dialect without a phonetic alphabet must be abandoned as unintelligible outside London) George Bernard Shaw, Pygmalion (1916) The reader of George Bernard Shaw's Pygmalion meets the surprising italicised passage in brackets above at the end of Eliza Doolittle’s third speech, and it is likely to provide relief, since the time and effort spent on the first three speeches of the famous flower girl are challenging to say the least, as the author acknowledges. Although dialects and accents had played a role in literature by authors such as Chaucer, Shakespeare or Dickens, their literary representation had always conformed to the written norm and stayed far from the real features of accent and dialect one could hear while walking on the streets. Accuracy on the scale illustrated by George Bernard Shaw had been avoided because, as his example illustrates, too much accuracy can compromise the reader's understanding of the character’s speech. Generally, therefore, dialects and accents in literary texts are literary re-‐creations that are only marginally concerned with accuracy. Even G. B. Shaw abandons his acknowledged failed attempt to reproduce Eliza’s dialect accurately in the remaining acts of the play. Authors take advantage of linguistic stereotypes easily recognised by the average reader, making sure that certain assumptions and images will be triggered and, consequently, help with the description of the characters in literary texts, who are differentiated, not only by what they say, but also by how they say it – to be more specific – by which linguistic variety they use. As soon as one linguistic variety is established as the standard variety in a society, (normally for political and social reasons), the others tend to be considered less prestigious and perhaps incorrect. Non-‐standard varieties easily become associated with peripheral geographic spaces and with a lower sociocultural status, as well as oral speechi, causing language to be an important element in the characterisation process. It is this same notion of correctness, propriety or adequacy of a language associated with a place or a person that has also made language a powerful element in the production of comic moments. Reading the play or seeing the film, it is hard to resist laughing at the inadequacy of Eliza Doolittle’s speech at Professor Higgins’ mother’s house, a place that, from its decoration and other characters’ attitudes, is interpreted as prestigious and thus a place where non-‐standard speech will seem “inappropriate”.
2
2
As we can see, the presence of literary dialects tends to provide added communicative and semiotic value to a text. In the context of translation, this means extra difficulties for the translator. He/she faces the difficult problem of having a linguistic variety in the source text which reflects the close relationship between the speaker, the medium and the context in which it is used. It evokes and explores extralinguistic knowledge, since speakers attribute value to the varieties, therefore establishing a hierarchy of varieties, as well as a hierarchy of speakers of those varieties. This creates moments of tension (Lane Mercier, 1995, 1997), “culture bumps” (Leppihalme, 1997), since the linguistic elements are both culturally conditioned and socially regulated, leading some authors to consider the translation of accents and dialects as impossible (House, 1973). The difficulty in translating literary dialects therefore lies not only in linguistic problems, but also in pragmatic and semiotic difficulties, since their presence in the text adds meaning far beyond the linguistic level. This is why it is important to discuss the translator’s decision to recreate or not to recreate linguistic variation and how he/she chooses to go about the problem, as this decision may modify, or even subvert, the work’s internal coherence. My purpose in this paper is threefold: firstly, to present for discussion a model which summarises the different strategiesii acknowledged by different researchers in several case studies; secondly, to present and discuss the strategies recognized in a corpus of 12 Portuguese translations of Bernard Shaw’s Pygmalion and Alan Jay Lerner’s My Fair Lady, as well as the contextual factors which might have led to them. Thirdly, and since this corpus comprises translations for different media – translations for the page, stage and screen – it is also the aim of this paper to observe whether it is possible to identify regularities between a specific medium and a certain strategy. Linguistic varieties in literary texts As mentioned above, the creative use of linguistic varieties in literary dialogue helps to inform the reader about which character is speaking and under which circumstances. It becomes a textual resource that helps the reader to define the sociocultural outline of the character, as well as his/her position in the sociocultural fictional context. Knowing the social stereotypes and assumptions readers have and share with the rest of the society they are a part of, the author uses fictional varieties with the expectation that this will encourage certain reactions and assumptions which will aid characterisation. It leads to the stratification of the participants in the dialogue, since the speakers tend to associate higher prestige with a standard variety, and, consequently, to undervalue other varieties
3
3
culturally associated with peripheral geographic spaces and with a lower sociocultural status. As established above, the literary recreation of accents and dialects has no pretensions of being accurate, but the degree of linguistic mimicry depends on the author's aesthetic, narrative, thematic, stylistic or functional objectives. The literary recreation of a linguistic variety is based on a previous selection, which results from different levels of mediation, being truly a pseudo-‐variety, a fiction, which Olga Brodovich has labelled as a ‘scenic dialect’ (1997: 26). Work by Blake (1981, 1995), Page ([1973] 1988) and Chapman (1994) suggests that the recreation of linguistic varieties results from the mediation, not only of the purposes to be fulfilled by the translation in the target culture, but also from what has been established by that system’s literary tradition or factors such as legibility, intelligibility and the medium in which the translation will come to light. The model Based on the conclusions of several descriptive case studiesiii as well as my own research, it is possible to summarize the different strategies for translating linguistic varieties in a model such as the one to be presented below. There are no prescriptive intentions behind this model or any kind of judgements on which strategies are preferred and which should be avoided. My aim has merely been to gather information from different case studies and organize it into a single typology of strategies, so that future researchers can more easily access this information and work with the same concepts. Existing case studies tend to refer to the same strategy in different ways or to classify the same concept as belonging to several different strategies. This typology is certainly not complete, and further research will surely bring more detailed information concerning these strategies or add more strategies, but, for now, it is hoped that it will bring the available information together more coherently. I shall now present the different levels of the typology, as well as the strategies it encompasses. Finally, I will discuss the strategies identified in a number of Portuguese translations, in order to try to establish what kinds of context encourage the choice of some of the strategies present in this model.
4
4
Figure 1: Strategies identified in the translation of linguistic variation in literary text
From the data available in several case studies, it seems the translator is first faced with the decision to preserve or not to preserve linguistic variation. A decision to use only one variety throughout the text will result in normalization (if the standard variety is used) or dialectization (if a non-‐standard variety is used) of the discourse. The common association made between standard variety and written discourse (cf. footnote 1) leads to a tendency to translate towards the standard variety, so it is not surprising that strategies of broad discourse standardization are often identified in contexts of authoritarian regimes or societies where there exists a broad gap between standard and non-‐standard varieties in terms of prestige (Hatim, 1990-‐91; Ben-‐Shahar, 1994). In the former case, state censorship is frequently the manner in which the standardization of the discourse is imposed; in the latter, the target public may express dissatisfaction with the use of non-‐standard varieties in such a prestigious medium as writing.
5
5
However, even though normalization strategies have been identified in several case studies (it was even formulated as one of the translation universals in Baker, 1993), it is not common to find cases where a complete normalization of the discourse has occurred. A broad normalization of the discourse tends to take place where the non-‐standard discourse appears in secondary characters’ speech. In addition, the relationship between the hierarchy established among plot elements and the recognition of linguistic variation as a peripheral element can also lead to a standardization strategy. If cases like Pygmalion, where non-‐standard speech plays a central role in the plot, make it difficult not to translate as non-‐standard speech, in other plays, Man and Superman also by Bernard Shaw, for example, where the non-‐standard speech only appears in secondary characters, the normalization of the speech could be easily accomplished without compromising the logic of the plot. As previously mentioned, the result will depend mostly on the importance and priority given to each of the elements of the plot. The opposite movement – the ‘dialectization’ of the discourse – can also be identified in specific historic moments. The studies developed by Brissett (1996), Cronin (1996), Findlay (2000) and Chapdelaine (2006) seem to support the link between nationalist movements and the strategy of dialectalization. Acting to ensure cultural and linguistic autonomy, nationalist movements see translation as “an act of reclaiming, of recentring of the identity, a re-‐territorialization operation” (Brisset, 1996: 165). The standard variety has now to compete with a non-‐standard variety, with which more prestige is associated, and is seen as the only one capable of expressing the identity of the group supporting it. It reveals a clear attitude in the subversion of the linguistic system, as well as the power relations it supports. Cases like the one reported by Annie Brissett in Quebéc or by Michael Cronin in Ireland, where translation into a minority and, consequently, non-‐ standard language has been used to affirm the language as the unique language of a certain nation, are well known among translation scholars today. The decision to preserve linguistic variation will mean a long process of choice amongst several other options. The translator will have to choose between preserving or not preserving the spatial and/or time coordinates of the source text, having four possibilities: » To maintain the space and time coordinates (e.g. Pygmalion in London in 1916) » To maintain the space but not the time coordinates (e.g. Pygmalion in London in 2007) » To not maintain the space and time coordinates (e.g. Pygmalion in Lisbon in 2007) » To maintain the time but not the space coordinates (e.g. Pygmalion in Lisbon in 1916)
6
6
The re-‐allocation of the plot can be seen as a tactic to reduce the “strangeness” effect caused by the use of a specific regional or social variety in a foreign environment. The placing of the plot in contemporary times might help the translator to avoid the difficulty of writing in a past-‐time variety of the target language. Keeping in mind the case of Pygmalion, it would not be difficult to understand the choice of any of these possibilities. The difficulty of finding a sociolect with the same social connotations as Cockney in 1916 would certainly make the options of re-‐allocating to Lisbon (if translating into Portuguese) or placing the plot in 2001very tempting. The former option would save the translation from the incongruence of having the characters speaking a Portuguese dialect in London. The latter, would allow the translator to avoid the problem of recreating a language from decades ago, about which he or she may not possess enough information. Nevertheless, and because of the modifications it imposes on the text, the non-‐preservation of one of the coordinates normally leads the public to classify the translator’s work as an adaptation. Having chosen to preserve the spatial coordinates of the source text (ST), the translator can then decide to make use of familiar features recognized as non-‐standard in the target culture or to make use of features not familiar to the target culture receiver. In the first case, several strategies can be identified, outlining a continuum from a more to a less normalized discourse: ·
Use of the standard variety in direct discourse followed by written indications informing the reader that the character was speaking in a non-‐standard variety. The introduction of explicit indications of the variety spoken by the character is a strategy of compensation for having erased the non-‐standard features.
·
Reduction of the linguistic variation to forms of address and honorifics. Even though it erases the non-‐standard features, this strategy allows the reader to understand the power relations between the characters
·
Upgrading the level of standard discourse formality. In this case, the linguistic variation comes from the fact that standard discourse is more formal in the target text when compared to the source text rather than from the use of non-‐standard features.
·
Usage of oral discourse features. Given the fact that oral discourse is less prestigious than written discourse, certain characteristic features of oral discourse are sometimes used to portray the discourse as non-‐standard.
·
Usage of features from different non-‐standard varieties. The strategy of not choosing a particular variety promotes the identification of the discourse as sub-‐standard and the polarization between standard/ correct/ higher status versus non-‐ standard/ incorrect/ lower status.
7
7
·
Usage of features of a specific non-‐standard variety. This strategy promotes the identification of a specific regional and/or social variety as well as the social meaning attached to it.
The use of features that are un-‐familiar to the target text receiver can be motivated by the fact that non-‐familiar features are also used in the source text (as for example in Anthony Burges’ A Clockwork Orange, where a new language based on English and Russian is created and shared by the members of a gang) or by the fact that the target system and public do not easily accept the idea of having foreign characters in foreign territory speaking national non-‐standard varieties. In certain cultural contexts, the suspension of linguistic disbeliefiv might not be activated in relation to non-‐standard varieties, even if it is normally activated in relation to standard variety. Following Olga Brodovich (1997), we understand that, in certain contexts, the translation of foreign linguistic varieties into some native varieties might not be well received by the target readership, who would not accept the incongruence of having an English character in London, the case of Eliza in Pygmalion, speaking a specific Russian dialect, for example. In this case, we can identify three different strategies: ·
The direct import of certain lexical features from the ST. This strategy consists on leaving some of the lexical items present in the target text (TT) un-‐translated. This seems to happen mainly when those lexical items were already foreign in the SC (one can think of Monsieur Poirot in Agatha Christie’s books where some French words are introduced to bring a ‘French tone’ to the character’s speech), or in cases where the vocabulary in the source language is familiar to the target text reader and will not compromise the understanding of the speech. In this latter case, the translator is the one choosing to introduce a ‘foreign tone’ to the text.
·
The introduction of lexical features from the ST, but following the spelling norms of the TT. Following Berezowski (1997), the polish translation of Herzog by S. Bellow seems to point to another strategy by means of which some source language lexical items are imported into the target text, albeit in a target language graphological form.
·
The conception of a ‘virtual dialect’. Cases like A Clockwork Orange by A. Burgess seem to force the translator to follow the author’s example and also create a new dialect based on the target language, but full of lexical items or syntactic constructions strange to the target text reader.
One could try to define all the possible circumstances underlying the choice of each one of these strategies, however, besides the difficulty of bearing in mind all the possible variables and contexts, one could only conclude that strategies cannot be seen as better or
8
8
worse, but as products of choices in context. In addition, and taking into consideration that the translators, often with no specialized linguistic knowledge of their own language, work normally with stereotyped features easily recognized by the target context community, we can in most cases recognize a combination of different strategies in the same translated text. The case study The corpus defined for this article comprises 12 Portuguese translations of George Bernard Shaw’s Pygmalion and Alan Jay Lerner’s My Fair Lady. BOOK
1966 1972
Pygmalion translated by Marina Prieto My Fair Lady translated by H. Silva Letra Pygmalion translated by F. Mello Moser Pygmalion translated by L. F. Rebelo and J. Palma e Carmo
1974
1987
Pigmalião translated by Mário Abreu
My Fair Lady translated by J. Nunes de Carvalho and Teresa Sustelo (RTP) Pygmalion translated by Ruth Saraiva (RTP) Pygmalion translated by Rosário Vieira (SIC) My Fair Lady translated by Eulália Ramos (SIC)
1994 1995 1996 2001
SCREEN
Pygmalion translated by António Lopes Ribeiro
1945 1961
STAGE
My Fair Lady translated by Filipe La Feria
My Fair Lady translated by Filipe La Feria
Figure 2: Portuguese translations of Pygmalion and My Fair Lady for page, stage and screen.
The use of a particular non-‐standard variety of British English – Cockney – is central to the plot of both plays, which can be summarized as the teaching of a standard variety by a phonetician (Professor Higgins) to a common flower girl speaker of Cockney (Eliza). Only Eliza’s lines were used and more specifically her lines as a speaker of Cockney, since it is not my purpose here to work on idiolect speech marks or the character's evolutionary progress. The analysis of the parallel corpus was made semi-‐automatically, using the computer programme, Systemic Coderv. A number of translation strategies were identified and will be discussed below.
9
9
Since one of the purposes of this paper is to take into consideration the variable ‘medium’, the analysis was designed to show which strategies were developed within each medium, what may have influenced these choices, and what can be learned about the differences between translating for the page, stage and screen. A summary of the main types of phenomenon observed is presented below. The presence of linguistic variation in the source text The source text, even though it resorts to stereotyped features, clearly promotes the identification of a specific regional and social variable – Cockney. The play can be divided in three parts, which correspond to the evolution of Eliza in her learning of Standard English. As discussed above, this case study will only focus on the first phase, where the effects of teaching cannot yet be noticed in Eliza’s speech. The high percentage of use of less prestigious varieties (social sub-‐standard variety and oral register) compared to the lower percentage of use of the standard variety, confirms the communicative meaning the nonstandard variety fulfils in the text – characterization of Eliza as an uneducated and low class flower girl in London. Concerning textual-‐linguistic features, the preference for graphic features instead of lexical or morpho-‐syntactic ones is noticeable, even though all the categories show a higher frequency than would be expected in the context of the English literary traditionvi. The target texts The first problem faced by the translator is the asymmetry of the Portuguese and English diasystems (dialect pattern). The translator of the 1972 translation, in an article about the difficulties of this translation, refers to the fact that there is no Portuguese low-‐urban-‐class variety corresponding to Cockneyvii. This situation led the translator to attempt to create a pseudo-‐variety resulting from the combination of features from different dialects generally depicted as “popular”, “uneducated” and sometimes “incorrect” by the community. But what did the other translators opt for? And to what extent has the medium influenced their decisions? Page translations All the translations published in book form show a preference for the preservation of the linguistic variation in the target texts (TT), as well as the time-‐space coordinates of the ST (the action always takes place in London of the beginning of the XXth Century). In what can be seen as consideration of the target context, we can also identify the use of features familiar to the target reader, even though different strategies were used.
10
1 0
All the translations make use of features generally perceived as distinguishing low educational level and low social status, but lower percentages are manifest in the translations before 1974 in relation to both the ST and the remaining translations. The year of 1974, which marks the end of the 48-‐years long right wing dictatorship, is a critical moment in Portuguese history and the end of Censorship is an important element to have into consideration in the context of translation of dialects as the translators would no longer face institutional pressure against the use of non-‐standard varieties in literary discourse. Bearing this in mind, it should be noticed that the translations published before 1974 show a higher discursive normalization. The translation of 1966 shows such a highly normalized discourse that the forms of address become an important additional element of distinguishing the participant’s social statusviii. It becomes also relevant to notice the fact that oral discourse features are used to distinguish the discourse as non-‐standard. Features such as ellipsis and contractionsix, playing with the association made between written discourse, orthographic norm and standard variety, seem to have been used to identify the discourse as non-‐standard. Even though this tendency is noticeable in all the translations, its frequency is lower in the pre-‐ 1974 translations. In terms of levels of discourse, there is a visible preference for lexical and graphic features (the so-‐called “eye-‐dialect”x) in loss of morpho-‐syntactic features. At this point, it is important to have in consideration the fact the, with the exception of swear words, lexical features are perceived by speakers of Portuguese as a less grave deviation from the standard when compared to graphic or morpho-‐syntactic features, i. e., non-‐ standard features in Portuguese can be organized in a continuum in terms of the level of gravity they are perceived to portray in relation to the standard variety. Lexical features would be closer to the standard followed by graphic features. Morpho-‐syntactic features would be on the end of the continuum representing the highest level of deviation from the standard discoursexi. Bearing this in mind, it is thus possible to assume that the preference for lexical features to be recognized in the translations before 1974 results in a higher normalization, as, even though distinguishing the discourse as non-‐standard, the use of lexical features would result in a target text’s discourse with a lower level of deviation from the standard discourse. The translations published after 1974 show a notorious preference for graphic features, the so-‐called “eye-‐dialect”. Even though still avoiding morpho-‐syntactic features, these translations assume a clear strategy of recreation of the linguistic variation characteristic of the source text and not one of normalization. Two of the translations (1962 and 1987) attempt to recreate a specific regional variety (the northern dialect), but they make use of only one feature -‐ the lack of distinction between [b] and [v] – and not very frequently and regularlyxii. As we come to understand
11
1 1
from the translator’s preface in the 1987 edition, the choice of this specific regional variety stems more from the fact that the translator grew up in that region and felt more comfortable using its linguistic features in his recreation than from the socio-‐semiotic correspondence between that variety and Cockney. Stage translations In the translations meant for theatre performance, we can also recognize the option to preserve linguistic variation, the time-‐space coordinates and the use of features familiar to the target audience. The choice of preservation of the time-‐space coordinates (and not of a re-‐allocation of the plot -‐ a strategy frequently used in theatre and present in one of the Brazilian translations of Pygmalion – Shaw, 1964) shows us an initial concern for adequacy vis-‐á-‐vis the ST. Similarly to the translations published in book form, all the theatre translations make use both of features generally perceived as identifying low educational level and low social status, and oral discourse features as non-‐standard features, but in higher percentages for both categories. The percentage of standard discourse is also very low, lower than in the published translations. In one of the translations (1945) the use of features from a specific regional variety are evident, but, once again, it shows a low frequency and makes use of a single feature -‐ the lack of distinction between [b] and [v] typical of the northern dialect. Concerning textual-‐ linguistic features, it should be noted that the higher percentage of morpho-‐syntactic marks almost non-‐existent in the published translations, as well as the use of several kinds of graphic marks such as for example the switch in vowel qualityxiii, monothongizationxiv, metathesisxv or nasalization of the vowel at the beginning of the wordxvi. Screen translations In the translations for subtitling it is possible to distinguish two groups with two different strategies. The first group is composed of translations from 1987 and 1994 for the public Television channel RTP, and the second group consists of translations from 1995 and 1996 for the private Television channel SIC. In the first group, the discourse has been highly normalized: non-‐standard features are absent and most ST features of oral discourse are omitted (ellipsis, interjections, contractions, etc.). The few non-‐standard features to be found here are lexical features and it could even be said that the social difference is only perceptible by the use of certain forms of address. In the second group, the use of both the features generally perceived as identifying low educational level and low social status is noticeable, as well as oral discourse features as nonstandard features. Even though morpho-‐syntactic features are almost non-‐existent, as in the published translations, a
12
1 2
higher percentage of lexical and graphical features is noticeable when compared to the rates in the first group. A commentary on the data From the analysis outlined above and among all the strategies identified, this paper will only focus on two main strategies: on the one hand, a strategy of normalization of the discourse; on the other hand a strategy of preservation of the social difference between the participants by the use of non-‐standard features and oral features. Normalization of the discourse As already seen, the tendency to normalize the discourse is noticeable in the translations published before 1974 and in subtitles broadcast on public Television. This suggests that the effort to keep the high level of written standard Portuguese might be due to the pressure of censorship in the case of published books, and by the state-‐owned RTP channel (which defines itself as public service) in the case of subtitles. The fact that these translations prefer to use lexical features than grammatical or graphical ones suggests that the translators are conscious of the importance of the sub-‐standard discourse in the plot, but want to follow standard spelling. The translation of 1972, despite making use of graphical features, always italicizes them in order to signal them as shifts from the norm. There is still another important factor which was certainly taken into account by the translators – legibility. With the exception of the 1966 translation, the other two translations were published in very popular collectionsxvii aiming at a middle class reader with less extensive reading experience and most probably not acquainted with the “eye-‐ dialect” tradition. The audiovisual text, addressing a very diverse public with different cultural backgrounds and reading skills, might not be easily read if constantly presenting graphic features, especially by the young (10 to15 years old) and the old (55 to 70 years old), who would be the target public of a film broadcast at 2 p.m. as Pygmalion and My Fair Lady were. In the case of subtitling, we can also consider the fact that the choice not to portray nonstandard features could also be a sign of acknowledgment of the product as audiovisual: relying on extra-‐linguistic elements, the translators may be relying on an appreciation of the participant’s different status on the basis of both visual (Eliza’s clothing) and audible input. Use of sub-‐standard features
13
1 3
The choice of sub-‐standard discourse might be interpreted as a consideration of accuracy with respect to the oral register of the source text as well as adequacy with respect to the target culture oral discourse. This is noticeable especially in theatre translations, since what is being produced is a written translation which will necessarily come to life in a different channel. In written translations, non-‐standard discourse can be suggested by the use of oral discourse features even though they would not be considered as nonstandard in Linguistics, simply because they represent a deviation from the written norm. On stage, the fact that it is, indeed, oral discourse makes those same features natural and expected, leading the translator to make use of nonstandard features more frequently. Even in relation to the influence of State censorship before 1974, one can see from the data that censorship seemed to accept non-‐standard oral discourse for the stage though it did not accept it in books designed for reading only. As mentioned above, in more recent book translations (1972, 1987 and 2001) and in private channel TV subtitling one can witness the use of what is called “eye-‐dialect”, which may be aiming at a better public reception, i.e. the translator tries to avoid a negative judgment from the public by using oral discourse features to suggest nonstandard discourse. In the case of subtitling, where the source and target texts appear simultaneously, translators can never escape the fact that there is always someone, nowadays a large majority of the viewers, who understands the source language, leading to the risk of what Gottlieb called the “feedback effect” (1994: 105). If the inclusion of oral or sub-‐standard features in writing can, in fact, lead to the risk of the translation being considered as a bad translation (Lefevere, 1992:70), the contrary is, nowadays, equally valid: viewers who understand the source text are normally very critical of subtitling that does not present specific discourse characteristics. The same could, in fact, be said in relation to the book translations, not only because this is a very well known play, but also because one of the translations is published in a bilingual edition. One can thus reach the conclusion that this strategy may be seen as an attempt to produce an accurate and adequate translation of the source text. In the case of subtitling, since this tendency has been noticed in subtitling aired by a private TV channel, the data seem to confirm the hypothesis suggested by Rosa (1999, 2001) that private companies may feel less responsible for upholding the standard. Conclusion and final remarks From this data we can conclude that different strategies can be found in the same translation and that the medium, even though important, is not the determinant factor in the choice of a specific strategy and exclusion of another. Nevertheless, it seems possible
14
1 4
to raise the hypothesis that the medium will determine the frequency and type of features used. If in the published translations several kinds of grapho-‐phonetic features are used, in the case of subtitling the apostrophe indicating the fall of a vowel is almost the only grapho-‐phonetic feature used, which confirms that the translator is well aware of not only the “strangeness” effect that this kind of feature will create, but also that it will influence the level of legibility. In theatre translations, not only are several kinds of non-‐standard features used, but they also demonstrate higher rates of frequency than book translations or subtitling – once on stage, the audience only has access to the target text and this can be an important aspect of focus on the main element of the plot, as well as a form of comic production (let us not forget that both plays are comedies). Further study needs to be done in order to understand the full extent of the influence of the medium in the translator’s decision-‐making process.
15
1 5
i As discussed by several authors, a common association is made between standard discourse and written
register and nonstandard discourse and oral register. Most of the time, the nonstandard varieties do not even have an established written norm, adding one more difficulty to the work of the translator. ii In this context I will use Kwiecinski’s definition of strategy: “a textually manifest, norm-‐governed, intersubjectively verifiable global choice of the degree in which to subscribe to source-‐culture or target-‐culture concepts norms and conventions” (2001: 120). A difference has to be established between strategies and procedures defined by the same author as “a result, textually-‐manifest, norm-‐governed and intersubjectively verifiable translational action applied to individual linguistic manifestations at a specific linguistic or textual level, e.g. the lexical level, the syntactic level, or the level of the generic structure (format)” (2001: 120). It is also the purpose of this paper to identify the different procedures, regarded as a consequence of the different strategies and possibly the norms motivating them. iii In the construction of this model several case studies were taken into consideration: Hatim (1990-‐91); Ben-‐ Shahar (1994); Dimitrova (1996, 2001); Brodovich (1997); Hatim and Mason (1997); Rosa (1999, 2001, 2004); Sanchez (1999); Leppihalme (2000a, 2000b); Hervey et al. (2001); Määtä (2004) and Nevalainen (2004). iv Following the term of “suspension of disbelief” coined by Samuel Taylor Coleridge (1987), we could use in this context the concept of “suspension of linguistic disbelief”. The concept of “suspension of disbelief” refers to the alleged willingness of a reader or viewer to accept as true the premises of a work of fiction, even if it seems to contradict everyday reality. The audience tacitly agrees to provisionally suspend their judgment in exchange for the promise of entertainment. In relation to this, the concept of “suspension of linguistic disbelief” would refer to the willingness of the reader to accept situations of the kind we often encounter in translated texts, where the characters speak a different language than the one spoken in the country where the plot take place. v For the purpose of this article version 4.5 for Windows of the computer program Systemic Coder, designed by Michael O’Donnell of WagSoft Linguistics software was used. The program is available on http://www.wagsoft.com/coder/. vi Several authors, such as Mills (1972), Blake (1981, 1995), Page ([1973] 1988) or Chapman (1994), take notice of the fact that the literary recreation of linguistic varieties is accomplished mostly through lexical features and at a very low rate of frequency (the majority of the features would appear at the beginning of the chapters to remind the reader of the character’s peculiar speech and in specific situations where the use of linguistic varieties would be most important). vii His exact words were: “De facto, na nossa língua não existe nenhum falar que seja tão comunicativo na sua difícil inteligibilidade, nem que seja do mesmo modo aceite como susceptível de acarretar discriminação social, ou pelo menos grave preconceito” (1984: 223). [In fact, in our [Portuguese] language there is no type of speech as difficult in its intelligibility or carrying the same connotations of social discrimination or prejudice (my translation)]. viii Example from My Fair Lady (transl. 1966) Elisa: Capitão, compre uma flor a uma pobre rapariga. (Oh Captain, buy a flower to poor girl.) (back translation) ix The concept of ellipsis refers to the elimination of part of the expression through a process of condensation, while the concept of contraction refers to the elimination of one or two vowels and/or consonant in a word. Example from My Fair Lady (transl. 2001) Elisa: Isso nem chega p’a uma violeta! (Isso nem chega para uma violeta!) (Standard Portuguese) x The term “eye-‐dialect” accounts for the introduction of changes in word spelling in order to portray the discourse as non-‐standard. Due to the association made between written register and standard discourse, some features of oral discourse like contractions, which work through graphical deviations in the spelling of words, might lead the reader to identify the speech as non-‐standard (Page, [1973] 1988; Walpole, 1974). xi This continuum was previously confirmed through a questionnaire organized by the author of this article in the context of her PhD thesis. xii Example from Pygmalion (transl. 1987) Elisa: Eu nem sabia como estaba bem. (Eu nem sabia como estava bem.) (Standard Portuguese)
16
1 6
xiii This concept refers to the substitution of one vowel for another.
Example from Pygmalion (transl. 1945): Elisa: Eu posso vinder as fulôres catrago no cesto. (Eu posso vender as flores que trago no cesto) (Standard Portuguese) xiv This concept refers to the substitution of a diphtong (two vowels in the same syllable) by one single vowel.
Example from Pygmalion (transl. 1945): Elisa: Eu cá quero ser caxêra duma loja de fulôres qu’há lá na’nha rua. (Eu cá quero ser caixeira numa loja de flores que há na minha rua.) (Standard Portuguese)
xv This concept refers to the alteration of the position or the letter in a Word.
Example from Pygmalion (transl. 1962): Elisa: Num foi pru male. (Não foi por mal) (Standard Portuguese)
xvi This concept refers to the situation when a non-‐nasal vowel in the standard variety becomes nasal in the
non-‐standard variety. This most usually happens at the beginning of the word. Example from Pygmalion (transl. 1974) Elisa: Eu não devia tê-‐lo comido; sou é muito inducada para o cuspir. (Eu não devia tê-‐lo comido; sou é muito educada para o cuspir) (Standard Portuguese) The opposite – denasalization of a nasal vowel – can also happen. Example from Pygmalion (transl. 1974) Elisa: E atão, que mal tem isso? (E então, que mal tem isso?) (Standard Portuguese) xvii The 1972 translation was first published in a low-‐price book collection sold with a newspaper (“colecção RTP/ Verbo”) and the 1987 translation was published in a low-‐price pocket-‐sized book collection very familiar to the Portuguese public (“Livros de Bolso Europa-‐América”).
Corpus: Source Texts: Shaw, Bernard, (1916) Pygmalion, New York: Brentano. ______ (1938), Pygmalion, [motion Picture of Anthony Asquith and Gabriel Pascal]. ______ (1957), Pygmalion, London: Longman [2000, 18ª edição]. ______ (1983), Pygmalion, [motion Picture for TV by Alan Cook]. Lerner, Alan Jay (1956), My Fair Lady, London: Penguin Books. _______ (1964), My Fair Lady, USA: Warner Brothers [motion picture of Gorge Cukor]. Target Texts: Book translations Shaw, Bernard (1962), Pigmalião [trans. Marina Prieto], Lisbon: Degree Thesis by the Faculty of Letters – University of Lisbon. Lerner, Alan Jay (1966) Minha Linda Senhora [trans. H. Silva Letra], Lisbon: Portugália. Shaw, Bernard (1972), Pigmalião [trans. F. Mello Moser], Lisbon: Edições verbo, Colecção RTP/Verbo. Shaw, Bernard (1984), Pigmalião [trans. Mário de Abreu], Lisbon: Publicações Europa-‐ América.
17
1 7
Lerner, Alan Jay (2001), Minha Linda Senhora, My Fair Lady [trans. Filipe La Feria], Lisbon: Publicações Europa-‐América. Stage translations Shaw, Bernard (1945), Pigmalião [trans. António Ribeiro Lopes], Lisbon: Theatre “Trindade”. Shaw, Bernard (1974), Pigmalião [trans. Luís Francisco Rebelo and José Palla e Carmo], Lisbon: Theatre “Teatro Maria Matos”. Lerner, Alan Jay (2001), Minha Linda Senhora, My Fair Lady [trans. Filipe La Feria], Lisbon: Theatre “Politeama”. Screen Translation Cukor, George (1987) My Fair Lady [trans. J. Nunes de Carvalho and subtitling by Teresa Sustelo for RTP]. Asquith, Anthony and Gabriel Pascal (1994) Pygmalion [trans. Ruth Saraiva for RTP] Cook, Alan (1995) Pygmalion [trans. Rosário Vieira for SIC]. Cukor, George (1996) My Fair Lady [trans. Eulália Ramos for SIC]. References: Ben-‐Shahar, Rina (1994), “Translating Literary Dialogue: a problem and its implications for translation into Hebrew”, Target 6:2, pp. 195-‐221. Blake (1981), Non-‐Standard Language in English Literature, Andre Deutch. Blake, N. F. (1995), “Speech and Writing: an Historical Overview”, The Yearbook of English Studies 25, pp.6-‐21. Brodovich, Olga (1997), “Translation Theory and Non-‐Standard Speech in Fiction”, Perspectives: Studies in Translatology 5:1, pp. 25-‐31. Chapman, Raymond (1994), Form of Speech in Victorian Fiction, London: Longman. Dimitrova (1997), “Translation of Dialect in Fictional Prose -‐ Vilhelm Moberg in Russian and English as a Case in Point” in Norm Variation and Change in Language. Proceeding of the Centenary Meeting of the Nyfilologiska Sällskapet Nedre Manilla 22-‐23 March 1996. Stockholm: Almquist & Wiksell International. ________ (2002), “Orality, literacy, reproduction of discourse and the translation of dialect” in Irmeli Helin (ed.) Dialektübersetzung und Dialekte im Multimedia, Frankfurt am Main: Peter Lang, pp. 121-‐139.
18
1 8
Gottlieb, Henrik (1992), “A New University Discipline” in Cay Dollerup and Anne Lodegaard (eds.) Teaching Translation and Interpreting: Training Talent and Experience Amsterdam: John Benjamins, pp. 161-‐170. Hatim, Basil (1990-‐91), “Intertextuality and Idiolect as Intended Meaning; A Concern for both translator and Literacy Critic Alike: with special reference to Arabic”, Paralléles: Cahiers de L’Ecole de Traduction et d’ Interprétation de l’ Université de Genève 12: Hiver, pp.77-‐87. Hervey Sándor et all. (2001), Thinking Spanish Translation – a Course in Translation Method: Spanish to English, London: Routledge. Kwiecinski, Piotr (2001), Disturbing Strangeness. Foreignisation and domestication in Translation Procedures in the Context of Cultural Asymmetry, Torun: Wydawnictwo Edytor. Leppihalme, Ritva (1997), Culture bumps: an empirical approach to the translation of allusions, Clevedon, Philadelphia: Multilingual Matters. _______ (2000a), “Päätalo Idioms and Catchphrases in Translation” in Erikoiskielet já käännösteoria: VAKKI-‐symposiumi XX. Vaasa: University of Vaasa. _______ (2000b), “The Two Faces of Standardization: on the translation of regionalisms in literary dialogue” in The Translator 6: 2, pp. 247-‐269. Määtä, Simo K. (2004), “Dialect and Point of View: the ideology of translation in the Sound of Fury in French”, Target 16: 2, pp. 319-‐339. Mercier,
Lane
(1995),
“Towards
writing/reading/(re)translating
a
Rhetorical
fictional
Practice
sociolects”,
of
RSSI:
Mimesis: Recherches
Sémiotiques/Semiotic Inquiry 15: 3, pp.105-‐121. _______ (1997), “Translating the Untranslatable: the translator’s aesthetics, ideological and political responsibility”, Target 9:1, pp. 43-‐68. Moser, Fernando de Melo (1984), Falares e idiolectos em Pygmalion de G. B. Shaw, como problemas de tradução, Tubingen: Gunter Narr. Nevalainen, Sampo (2004), “Colloquialisms in Translated Text. Double illusion?”, Across Languages and Cultures 5:1, pp. 67-‐88. Page, Norman [1973] (1988), Speech in the English Novel, London: MacMillan Press. Rosa (1999), “The Centre and the Edges. Linguistic Variation and Subtitling Pygmalion into Portuguese” in Jeroen Vandaele (ed.), Translation and the (Re)Location of Meaning, Leuven: Leuven Research Center for Translation, Selected Papers of CETRA, 1994-‐1996, pp.317-‐337.
19
1 9
______ (2001), “Features of Oral and Written Communication in Subtitling” in Yves Gambier and Henrich Gottlieb (eds.), (Multi) media translation: concepts, practices, and research, Amsterdam: John Benjamins, pp.215-‐221. ______ (2004), Tradução, Poder e Ideologia: Retórica Interpessoal no Diálogo Narrativo Dickensiano em Português (1950-‐1999) Ph.D. diss., University of Lisbon. Sanchez, María (1999), “Translation as a(n) (Im)possible Task: Dialect in Literature”, Babel 45: 5, pp. 301-‐310. Shaw, Bernard (1964), Santa Joana e Pigmalião [trans. Dinah Silveira de Queiroz, Miroel Silveira e Fausto Cunha], Rio de Janeiro: Delta.
20
2 0