SLANG AND DIALECT IN HARRY POTTER: THE TRANSLATOR AS MEDIATOR OF MARKED LANGUAGE

The present research covers the intersection between Descriptive Translation Studies and Children’s Literature and aims to investigate the important role of the translator as mediator of marked language, more specifically by dialect and slang. This paper is based both on Santos’s (2010) concluded mastering thesis which investigated the translation of the dialect of the character Rubeus Hagrid and Santos’s ongoing PhD research which deals with translation into Brazilian Portuguese of slang words in the Harry Potter series. The initial hypothesis is that dialects and slang words will be translated by standard language and this translator choice is influenced by the particularities of its readership.


Caroline Reis Vieira Santos holds a master's degree in Translation Studies by
Universidade Federal de Santa Catarina and develops an ongoing PhD research in the same Postgraduate Program. Her areas of interest are Descriptive Translation Studies, the Translation of Children's Literature, the Translation of Dialects and Slang and Text Register. Currently she has a CAPES scholarship.

First remarks
The aim of this paper is to show the importance of the translator's role as mediator of marked language or non-standard language such as dialect and slang. Both slang and dialect may be a translation challenge even for the most experienced translators due to its intricacy with cultural factors. The issue is even more delicate when we are talking about literature written for children due to the particularities of that readership: the limited world knowledge (related to the short life experience of these individuals); the ongoing development of reading competence; and the conflict between standard language learnt at school and non-standard language variety.
In addition to all mentioned above, the translator's attempt to reproduce a dialect or certain slang has its pitfalls: the risk of sounding artificial in the target language, the risk of causing an undesired effect such as comicality or the risk of producing something that will not achieve the communicative goal intended. Yet, the translation of marked language for standard language can disfigure a character's speech or a group speech. How, then, can a translator solve this problem? This paper will show the translation method 2 adopted by two translators in the case of dialect and an initial hypothesis based on a pilot study in the case of slang words.
To fulfill its purpose, this article will be organized as follows: (i) the presentation of two authors' considerations on the topic of marked language; (ii) the presentation of the findings of Santos (2010) mastering thesis; (iii) the presentation of the initial hypothesis of Santos's ongoing PhD research based on a pilot study done; (iv) the discussion of the role of translator as a mediator in the case of dialect and slang present in children's literature; and (v) final remarks.
1 Marked language: Krings (1986), Klingberg (1986) and Jentsch (2006) Both Krings and Klingberg published their books in the 80's, in the beginning of the consolidation of Translation Studies as a discipline. Moreover, there are not many resemblances between the two works despite the fact that both have mentioned the translation of marked language: Krings (1986) was interested in translator's education and Klingberg (1986) has published a pioneer study on children's literature translated. Unfortunately, until the present moment not much literature about slang or the translation of slang was found. That is the reason why they were situated under the more generic label of marked language adopted by Klingberg (1986) and Krings (1986).
Talking from the translator's education background, Krings (1986: 273) points out that "[…] the most frequent case of reduction strategy consisted in giving up the markedness or the metaphorical character of a source-language text item and replacing it by a nonmarked or a non-metaphorical equivalent." The author has made a study with eight subjects asking them to translate a certain text into the native language and from the native language into a foreign language. In that study Krings (1986) has found that reduction strategy was adopted when the subjects faced marked language. Klingberg (1986) has investigated many possible translation problems in children's literature, one of them being specifically the translation of dialects and non-standard language. According to the author, there are two opinions on how to handle dialect: "One holds that, because of the difficulties, a dialect of source-text should not be translated as dialect of the target text. The other opinion emphasizes that the function of a dialect in a source-text and wants it preserved in some way" (Klingberg, 1986: 70-71). Klingberg does not make clear whose difficulties he is referring to, if these difficulties are related to the source and target text language system, if they are related to the readership needs or if they are related to the intricacy of dialect and cultural aspects. However, in all the material analyzed by the author, the dialect was translated by standard target language showing that, although there are two different opinions, there seems to be unanimity concerning the use of what he calls standardization strategy.
In the same direction, Jentsch (2006) has found in her study that has explored many topics concerning the translation of Harry Potter book series, including the translation of Hagrid's dialect, that in the three translations analyzed by her (one into Spanish, one into French and one into German), Hagrid's dialect was translated as standard language with few oral and informal marks. Nevertheless, the author shows a prescriptive tone by saying that the translators should have chosen a harmless solution to render Hagrid's dialect. She does not explain, however, what is understood by harmless solution nor gives an example of how that can be done.
As one can see, there are not many studies on translation of marked language and they do not investigate the topic very deeply. Santos's (2010) mastering thesis in Translation Studies main purpose was to investigate the "preferred or recurring patterns of linguistic behavior, rather than individual or one-off instances of intervention." (Baker, 2000: 245) of the Brazilian translator Lia Wyler and the Portuguese translator Isabel Fraga concerning Hagrid's dialect representation in the first Harry Potter book -Harry Potter and the Philosopher's Stone (Rowling, 1997). After observing these strategies adopted by the translators, they were compared to what Krings (1996) calls reduction strategy and Klingberg (1986) considers standardization. For the purpose of this study, a dialect is understood as a variant that you speak because you 'belong to' (come from, or have chosen to move into) a particular region, social class, caste, generation, age group, sex group, or other relevant group within the community. (Not all of these are relevant at any one time or place: but the combination of them may be.) [...] (Halliday, 1989: 44) In the specific case of the character of the book, what we have is not actually a real dialect, but a representation of the Somerset dialect. Rowling has printed in the character's speech some dialectal marks probably based on her speaker knowledge. The recognition of the character's dialect was possible due to comparison between some non-standard marks present in Hagrid's speech and Peter Trudgill's (1999) study of English dialects. In fact, it was an interview given by Joanne Rowling and Stephen Fry broadcasted by BBC and transcribed by the fan website MuggleNet 3 (BBC Radio 4, 2005) that gave the hint that Hagrid's dialect was Somerset.

Method
The first step of data analysis consisted in reading all Hagrid's utterances in the source-text hardcopy and observe the dialectal marks used only in his speech and that did not belong to the standard language. After recognizing these marks, they were gathered in some categories, that will be detailed in 3.1.2, created to cover the phenomena arisen from the corpus.
After observing the hardcopy, a parallel corpus with the source text and the two translations was created. The source-text and the Brazilian translation were already available in electronic text format and aligned and ready to be used because they were part of PEPCo (Fernandes;Bartholamei, 2009), which no longer exists because the researcher responsible for it has decided to close it. This corpus was very similar to Imao's (2008) Casual Conc Simple Web Parallel Concordancer.

Corpus compilation
The European Portuguese translation of Harry Potter and the Philosopher's Stone was captured (the text was scanned from hardcopy into the readable PDF), proofread (the scanner does not read some characters in a proper way, replacing, for example, the letter 'I' for number '1'), edited (some tags were added, the text was aligned in a parallel form) and then converted into .txt format using Notepad++ (Don, 2011) and added to Portuguese English Parallel Corpus (PEPCo) (Fernandes;Bartholamei, 2009). After finishing corpus compilation, described above, all Hagrid's utterances were selected in both source and target texts and separated from the rest of the text.

Categories of analysis
After finishing compiling the corpus, tags were added indicating the presence of dialectal marks (in the original) and oral marks (in the target texts). The tags applied were the following:

Chart 1 -Categories of analysis text text
After applying the tags, the next step was aligning the source and the target texts for data analysis using 'keyword' tool provided by PEPCo. The searches using the tool 'keyword' were done using the tags described above as keywords. The results of that search will be shown in the next section.

Data Analysis and Discussion
By doing searches using the tags as keywords it was possible to find the number of occurrences of each dialectal (in the case of sourcetext) and oral (in the case of target texts) marks in the corpus. This quantitative analysis supported a qualitative analysis, considered most fruitful for the purposes of this research. Corpora tools provide us with an empirical view of translation and subsidize a qualitative analysis that interpret the data emerged from corpus observation.
The data analysis has shown that both translators have neutralized dialectal marks from Hagrid's speech through the usage of standard language. One of the possible reasons for the use of standard language may be the difficulty in reproducing or recreating the group of specific variables related to the individual's background as geographic location, gender, religion, caste, generation, as says Halliday (1989), in the target text. Nevertheless, the translators Lia Wyler and Isabel Fraga have used oral marks transcribed to the written medium in attempt to avoid the total neutralization of Hagrid's dialectal marks. However, the Portuguese translator has opted for using more oral marks in Hagrid's speech in comparison to the Brazilian translator (Santos, 2010), as the table below shows: It is important to keep in mind the difference between dialectal and oral marks. Just a manifestation of a group of variables can be considered a dialectal mark, thus they are only present in the original. The occurrences in both target texts were considered oral marks because they are not related to or are a manifestation of a group of specific variables such as gender, religion, age group, generation etc. These oral marks could be present in the speech of any person whereas the dialectal marks used in the original convey the idea that an individual shares the same background than other individuals of a community, expressed through dialectal variables.
The differences between the translation methods adopted by both translators possibly reflect the differences between Brazilian and Portuguese readership, publishing norms in each country and censorship. The translator, thus, has a very important role as a mediator in the translation process: one has the commitment to the original text but has also to bear in mind that that text has to fulfill a communicative role in the target audience. Depending on the translation choices, the book can be successful or not, can reach its target audience with the same or similar function and characteristics.
To conclude this short presentation of the results of Santos's (2010) mastering research, it is important to mention that the choices made by the translators do not depend only on their will. The translation activity is related to different norms, mainly the translation of children's literature: (i) source-related norms; (ii) literary, aesthetic norms; (iii) business norms; (iv) didactic norms; (v) pedagogical norms; and (v) technical norms (Rudvin;Orlati, 2006). These norms many times are prevalent over translators' decisions and sometimes they also contradict each other.

Slang translation in the Harry Potter series
The proposal of Santos's ongoing PhD research is also to analyze the translation practice into Brazilian Portuguese of slang words in the seven books of the Harry Potter series by Lia Wyler. The focus is also on how the translation methods chosen by the translator are going to influence text register.
This study considers slang as "words or combination of words that have a paradigm and do not show a crystallized syntactic order (what makes them different from expressions), which have ephemeral existence and veiled meaning and belong to the oral mode (but can be transcribed) and are not used in standard language, but rather in situations where there is social interaction and group identification" (Dubois, 2009;Eble, 1996;Gurgel, 2009;Halliday, 1989;Preti, 2000;Xatara;Falcão, 2005). Eble (1996) also draws attention to what she calls 'national slang' […] with the possibility of instant and widespread communication in recent years, the group-identifying functions of slang for the population at large may have been diminishing. Instead, speakers may be using slang to identify with a style or an attitude rather than a group. (Eble, 1996: 119) Although Eble makes this differentiation, one cannot deny that even though the group idea is weakened, there is still some kind of identification, an affiliation between society members based on their preferences, hobbies etc. Bearing in mind that Rowling's readership was not intended to be only a certain group of children and teenagers, but young readers from Britain in general, it is much more likely that she had used in her books what Eble calls 'national slang', but restricted to a bigger and more comprehensive group: the group of young people.
Amongst the characteristics related to slang use, one is of great interest to this study: "An important factor governing the use of slang items in a particular conversation may well be register -the varieties of language used by a single speaker in response to different occasions." (Eble, 1996: 112). According to Hatim and Mason (1990) that applied the notion of register specifically for translation purposes, text register is language variety related to use in the various contexts in which it takes place; the type of language selected will vary according to the situation. The register is divided in three main characteristics related to the language use: mode, tenor and field. Roughly speaking, the mode is related to the medium (written or spoken and the variations that exist between the two modes: written to be spoken, recitation etc.); the tenor is the relation established between the communication participants (more or less formal, more or less hierarchical etc.); and field is the "combination of subject matter, in so far as it influences such parameter as formality, complexity and presentational modes" (Fawcet, 1997: 80). The initial hypothesis of this study is that the translation of slang words for non-slang words will strongly affect the tenor of the discourse through the increasing of formality and distance between the participants. The other two components of register, mode and field, will also be affected, but in a smaller degree.
Defining '(in)formality' is a challenging and tricky task, but a working definition must be adopted. For the purpose of this research, the 'formal/informal' status is attached to social context in which a word is first used with a new meaning (Neufeldt, 1999). An example is the use of the words 'mouse', used as computer technical term, and the word 'chick' used by young people to refer to a girl or woman, being the latter considered informal.
Section 4.1 above will describe the method used in a pilot study with Harry Potter and the Philosopher's Stone into Brazilian Portuguese.

Method
The method presented in this section will probably be used to investigate the seven books of the Harry Potter book series. Apart from the first three books, that are already in digital format, corrected and aligned in the paragraph level 4 , the four remaining books from the Harry Potter series still have to be captured into rich text format (.rtf), corrected and edited.
For the moment, the results presented here were from a pilot study made with the first book of the series, Harry Potter and the Philosopher's Stone. This pilot study consisted of a manual observation of the source-text hardcopy in which some slang words candidates were found. A slang dictionary (Ayto; Simpson, 2008) was used to confirm if the slang words candidates were really considered slang and to look for their meaning. Chart 3 presents the slang words found in the source-text of this pilot study, the definition brought by Ayto and Simpson (2008), the name of character in whose speech the slang word was found and the function they have in the context in which there were found.

Chart 3 -Identification and classification of slang words in the source-text
After defining the source-text slang words list, this list was used as a starting point for the parallel analysis. These slang words were used as keywords for a parallel search using Imao's (2008) Simple Parallel Concordancer, a script for parallel concordance 5 . In this Parallel Concordancer the researcher copy and paste in two different boxes both the source and target text in text format formerly aligned. The analysis was done by using the keyword-incontext tool and the output form used was 'table'. The results from the parallel analysis are displayed in chart 4.

Chart 4 -Parallel analysis of source and target text in Harry
Potter and the Philosopher's Stone As one can see, one slang word from the source-text was translated by different words in the target-text. To confirm if the words used in the target-text were also slang, three different dictionaries were used: a slang dictionary (Gurgel, 2009), an on-line open-source dictionary of informal words (Dicionário Informal, 2012) and a common lexicon digital6 dictionary (Ferreira, 2008). Chart 5 displays a summary of the criteria used to analyze the translations of the slang words: To become Verb miniAurélio 6. Digital dictionary refers here to a CD-ROM that can be installed in a computer.

Chart 5 -Analysis of the terms used in the Brazilian Portuguese translation of slang words
These criteria were taken into consideration due to its potential influence on the text register. The method used in this pilot study has its limitations, once it is really time consuming and labor intensive to analyze manually each of the seven books. It is still necessary to develop a method that can be applied to the lager corpus of this research project. More details about the cross-checking between the slang list and the corpus still have to be established. A possible option is to use the program Windows Diff (Allen, 1999) to cross-check the slang words list based on Ayto and Sympson (2008) and the sourcetexts, but further tests using that program need to be done.
There is also another methodological problem to be solved regarding specialized slang dictionaries in Brazilian Portuguese. Until this moment, it was not possible to find a dictionary elaborated rigoursly by lexicographers or philologists that covers slang words that can be used as a parameter for defining if a word can be considered slang or not. Common lexicon dictionaries many times show criteria inconsistencies in categorizing slang and different dictionaries classify the same words under different labels.

Initial Analysis and Discussion
According to a first analysis and the data shown in charts 3, 4 and 5, all the translations of slang words have maintained the same communicative function and a same slang has been translated by different words in the target text, but not all of them were slang. Moreover, the informality has been maintained, even though there has been a tendency to standardization (Klingberg, 1986) or reduction (Krings, 1986) every time a slang word was translated into a nonslang word.
To summarize, the Brazilian translation practice regarding the slang words present in the first book of Harry Potter's series is, in most cases, to render slang words for common lexicon, but maintain a certain degree of informality. That practice influenced text register (i) through the increasing of formality in text register; (ii) through the decrease in the level of hierarchy between the characters' speech, once without the presence of slang words, the differences between the utterances is less evident; (iii) through the attenuation of social groups cohesiveness; (iv) through the smaller differentiation between moments of social interaction and intimate moments; and (v) through a smaller contrast between adults that belong to a higher hierarchical position -as professors and parents in the story -and other adults that occupy a lower hierarchical position, such as Hagrid, who identifies more with teenagers than with adults.

Translator as mediator of marked language: slang and dialect
The main purpose of this paper was to show the importance of the translator's role in two specific topics in the translation of children's literature: dialect and slang. Both topics constitute a real challenge to translators due to their intricacy with cultural aspects expressed through language marks. Without the mediation of the translator, the target text might not have achieved its communicative goal or could be rejected by its target readership.
The differences between Adult and Children's Literature cannot be neglected. As Shavit (2003) says, if we use the same parameters used in Adult Literature to criticize Children's, we will be like Cinderella's sisters, "[…] who cut off their toes and heels respectively, we would be cutting off our noses to spite our faces if we try to wear the shoe; we shall have achieved little, and be handicapped to boot." That means that the translator's interference in Adult Literature is different and must be seen differently from the interference in Children's books. This does not mean that one should underestimate the youngster's cognitive and interpretative skills, but, instead, respect the development phase that the individual is going through. Thus, the criticism applied to Adult Literature cannot be the same applied to Children's Literature due to the differences between the readerships, the editorial market, the censorship (applied mainly to the latter) and the translational norms.
This implies that the translator's interference must not be seen as a negative aspect of the translated text, but, in many instances, as a positive aspect. The translator knows one's readership better than the source-text author, and what may be perfectly acceptable for the source audience, may not be good for the target audience. This requires a great responsibility from the translator of working not only as a linguistic mediator but also as a cultural mediator. Many times the translator faces cultural items or marks in the text and one is responsible for making the child reader get in touch with cultural differences and different language registers when, at the same time, language itself cannot be a barrier. In that case, the translator has to decide to what extent s/he thinks that one's reader will be able to deal with and understand different cultural marks. This choice, as any choice one makes, is open to criticism. Despite that, the challenge must be solved, otherwise there will be no translated text.
In the first example, the one that has investigated the translation of Hagrid's dialect into Brazilian and European Portuguese, regardless the fact that dialectal marks have been drastically reduced in Hagrid's speech, the maintenance of these marks might have given the wrong idea of Hagrid's origins (for example, supposing that the translator had chosen to translate Hagrid's speech using a country dialect, this may cause in Brazil an undesired comicality due to its artificiality), or could threaten the reading skills of the younger reader that may not understand the translator's attempt to print dialectal marks in the written mode. The option for trying to render source dialectal marks for target dialectal marks could even be censured by the editors, who are preoccupied with parents' and educators' opinion concerning nonstandard language manifestations. In addition to all these possible reasons, there is also a great difficulty in trying to translate linguistic marks that reflect a cultural and social background and the scarce time professional translators have to do researches.
As Hatim and Mason (1990: 4) point out: "Like producers and directors, translators have to be constantly alert to the social implications of their decisions. The representation in a ST of a particular dialect creates an inescapable problem: which TL dialect to use?" The answer to this question, based on the practice of five different translators to four different languages (Brazilian Portuguese, European Portuguese, German, French and Spanish 7 ), is that no dialect is used, but instead, some oral marks are added to the characters' speech in order to differentiate it from other speeches. These oral marks have been used in greater or smaller extent by each translator, but all of them have made an attempt not to completely erase Hagrid's differentiation. In the study done by Santos (2010) as shown above, the Portuguese translator has printed more oral marks than the Brazilian.
Concerning slang words, the same remains true. Social groups are unique and are bounded by specific cultural characteristics. However, in the case of slang words, the problem might be a little less difficult to solve. Even though slang words are generally ephemeral, they are more independent of social factors in comparison to dialects, and do not depend on a group of characteristics (in opposition to only one) in the same way as dialects do. That is the reason why it may be easier to find a target lexical item with a similar function or meaning in a similar group of the source-text.
As slang words are more systematized in dictionaries in relation to dialects, it may be easier for the translator to use reference books to help her or him to render the source-text slang for a similar in form or function target text slang. The initial hypothesis, based on the pilot study presented, is that the Brazilian translator has used similar translation method for both dialect and slang words in the sense that she will render the original slang words for less marked lexical items. By doing so, the translator will try to show her target reader the existence of different social groups in the books that use language to guarantee identification at the same time that guarantee the exclusion of those who do not make part of that group, and, consequently, are not familiarized with the veiled vocabulary shared only by the participants.
The translator's role, then, goes beyond that of simply translating textual structures. One is going to contribute to the reader's understanding and to the success of a book. One is responsible for presenting cultural references and cultural differences many times expressed through grammatical and lexical patterns. The human 7.These last three translations were analyzed by Jentsch (2006). factor, in this case, is irreplaceable and of great importance for the increasing world knowledge of the young reader: literature is many times the opening door for knowing the unknown and learning how to deal with differences. A non aware translator may jeopardize not only the reading acquisition through the creation of linguistic barriers, but also the access to other cultures and literary values.
There is not an ideal strategy 8 that can be used to all text genres or to all readerships that may help the translator to deal with one's problems. The translator's interference is crucial to decide what the best option is taking into consideration the readership and the translation norms at stake.

Final remarks
To conclude this brief paper, it is important to mention that Children's Literature requires much responsibility of the translator due the particularities of its readership. One's mediator role is even more important due to the fact that the youngster world knowledge is limited by his or her short life experience. The intellectual maturity is also in development during the school years, what may require that the translator works as a mediator to show and explain cultural references from the source-text to which the reader may not be familiarized.
It is also important to bear in mind that this short presentation of Santos's thesis and dissertation has left many details aside, showing just a small part of a long work that has tried really hard to be less biased as possible and not to carry many pre-conceptions about the translated texts.