The author thanks the National Council for Scientific and Technological Development - CNPq (PQ-1D 305063/2014-6) for funding received.

1. The distinctive nature of the language of translation

1The distinctive nature of the language of translation and the intimate relationship between language and culture can be fruitfully investigated through the use of a parallel corpus as a resource for teaching/learning various features of the translated text. In recent years, research into corpus-based translator training has received considerable attention in the literature, particularly with respect to the use of corpora for pedagogic purposes (Hatim 2001; Granger et al. eds. 2003; Zanettin et al. eds. 2003; Laviosa 2008, 2009).

2In this learner-centred context, corpora are designed and used for retrieving translation equivalents as well as for improving students’ understanding of actual translation regularities (Laviosa 2008, 2011; Camargo 2011). Another important use of corpora of translated text is that they allow the analysis of variation in the translators’ output (Baker 2000).

3Corpus-based translation studies have also made an important contribution to translation theory and practice in describing how translators use the target language. The use of corpora can increase students’ awareness of the distinctive nature of translation as a communicative event which is shaped by its own goals, pressures and context production. Even though Baker’s work is based on comparisons of translated and non-translated texts of the same language, she suggests that,

any patterns we might similarly identify as distinctive on the basis of examining a translator’s output should next be compared directly with the source text in order to address the question of the potential influence of the source language and/or author style. (Baker 2000, p. 255)

4After outlining the main features of the language of translation, as put by Baker (1996) in corpus-based translation studies, and introducing the theoretical source for examining translation options for cultural markers, as elaborated by Nida (1945) and Aubert (2006) on cultural domains, as well as presenting the student-centred teaching approach, as adopted by Johns (1991a, 1991b) and Laviosa (2008, 2010), this paper illustrates how the professionally-oriented corpus‑based translation classroom can benefit from combining the three proposals for discovering linguistic regularities and the distinctive features of translational behaviour alike. Also, the proposal for adopting a three-phase methodological procedure draws on Laviosa (2008), which reports on a case study of corpus-based translation pedagogy.

5Investigations carried out by Baker (1996, 2000) at the Centre for Translation and Intercultural Studies (CTIS), housed at The University of Manchester, detected certain characteristics which typically occur in translation (Baker 1996, pp. 180-184). Some of these patterns were explained in terms of simplification, viewed as a tendency on the part of translators to simplify the language or message or both (Baker 1996, 2000). One of the features which may contribute to simplification is change in punctuation in order to simplify and clarify the translated text (Laviosa 1996).

6Explicitation tends to occur when implicit elements in the source text, due to information previously mentioned or inferred from the context, become explicit in the target text, in order to favour clarity (Baker 1996, 2004). A number of textual practices include the use of supplementary explanatory phrases, the spelling out of implicatures, the insertion of connectives to improve the logical flow, and hence the readability of a text and the resolution of ambiguity (Scott 1998).

7Levelling out refers to the tendency for translations to “steer a middle course between any two extremes” (Baker 1996, p. 184). This suggests that translations are, on the whole, far more similar to each other than to the texts of corpora of original compositions.

8Normalisation may be defined as a tendency to exaggerate features of the target language and to conform to its typical patterns (Baker 1996; Scott 1998). It may be observed in translated texts both in the use of individual words and in the use of typical grammatical structures, punctuation and collocational patterns or clichés. Long and elaborate sentences as well as redundant elements in the source text are replaced by shorter structures, and redundancies are often omitted. Usually the rhythm is more fluent, since unusual punctuation in the source text is standardised so that the translated version runs smoothly. Kenny (2001, p. 66) shares Baker’s ideas, and points out that translators tend to use more conventional solutions related to uncommon expressions in the source text. Also, Berber Sardinha (2002, p. 18) reports that, in normalisation, there is a minimisation of creative or less common aspects of the source language use. The analysis of lexical choices in the original work vis-à-vis the translator’s choices may reveal aspects of normalisation if, for example, it is shown that more creative choices in the source text have been rendered by other less marked options in translation (Ibid.). This feature is possibly influenced by the status of the original work and source language, so that the higher the status of target language the higher the tendency towards normalisation.

9There are several features which contribute to text normalisation in translated texts, which may involve aspects of simplification and explicitation. In their corpus analysis, students may deal with features involving, for instance, changes in the sentence length between the two texts, due to factors related to differences in punctuation, and explicitation of ellipsis. Also, students may examine occurrences of omission and register shift from the colloquial language which is typical of some characters in the source text to more formal language in the translation.

2. Linguistic cultural markers

10Cultural elements referring to specific realities of the source language universe are viewed as cultural markers by a number of translation scholars (Nida, 1945; Corrêa 2003; Aubert 2006; Camargo 2005). However, their identification and classification may frequently offer challenges to the analyst. Aubert explains that,

the cultural marker is not perceptible in its linguistic expression taken in isolation, nor it is confined to its original discursive universe. The cultural marker only becomes visible (and, therefore, is updated) if this original discourse (a) incorporates a differentiation in itself, or (b) is placed in a situation that highlights the differentiation. (Aubert 2006, p. 25)

11The approach and procedures used for examining translation options for cultural markers draws on Nida’s proposal on cultural domains (Nida 1945), as well as on the reformulation of definitions of cultural domains suggested by Aubert (2006), as they were found to be more precise during corpus analysis. Aubert (2006) distinguishes four cultural domains:

• ecological domain: words designating human beings, animals and objects in their natural condition or used but unchanged by men. Here are some examples of cultural markers taken from João Ubaldo Ribeiro’s Viva o Povo Brasileiro and its respective translation An Invincible Memory, which students may identify from corpus observation: aipim/“cassava”; caatinga/“wilds”; cacharréo/“cacharréo” [male whale]; mutuca/“motuca flies”; pé de acácia/“acacia tree”.

• material domain: words involving man-made objects or human activities. Examples: agogô/“tin bell”; caleça/“carriage”; gamela/“bowlful”; moqueca/“miscellany of fish”, “combination of fish”; saveiro/“boat”.

• social domain: words describing man, his social and professional classes and functions, origins, hierarchical relationships, as well as activities and events which establish, maintain or transform these relations, including linguistic activities. Examples: afoxés/“black band”; mestre de terra/“shore chief”; moleque/“black ass”, “black bastard”, “brat”, “scamp”, “street boy”, “pickaninny”; cafetina/“madam”; mulher-dama/“woman of the profession”.

• ideological domain: words designating beliefs, mythological systems, and spiritual entities associated with these systems, as well as the activities and events generated by such entities. Examples: búzios/“seashells”; Ogum/“Ogoon” [spiritual lord of tools and iron]; mãe de santo/“great mother”, mandinga/“sorcery’; mandingueira/“conjure”.

12Depending on the context in which they appear, cultural linguistic markers may be classified in one or more domains, since some cultural terms may have different meanings and connotations. For instance, “capoeira” may be related to the ecological domain, when referring to vegetation, forest, or to the material domain, when designating an art or a battle-ready fighting.

3. Corpus and data-driven learning

13Over the past decade, we have witnessed what Cook (2010) has named “the revival of translation” in foreign language learning and teaching. Having taken into consideration the principal reasons for this remarkable comeback, Laviosa (2011) entered a typical undergraduate language classroom in Europe and found out the students’ opinion that translation is an enjoyable and challenging activity, which boosts confidence in the use of the second language (L2). Also, they told her it is a skill which can be very useful in today’s increasingly globalised work environment. Laviosa (Ibid.) claims that translation develops accuracy, fluency, language awareness and intercultural competence, and explains that, if we cast an eye over the surrounding interdisciplinary landscape, we will find other compelling reasons for the readmission of translation in the language classroom, such as: (a) cognitive studies in second-language acquisition (SLA) have shown that the different languages of an L2 user are interconnected and are part of one dynamic, integrated system; (b) in bilingual education, translation is advocated as one of the bilingual instructional strategies in which two or more languages are used alongside each other; (c) in foreign language education, translation, together with interpretation, is considered to be one of the functional language abilities which enables learners to achieve translingual and transcultural competence; and (d) translation studies has unveiled the nature of translating as a form of linguistic mediation and a complex social, cultural, textual, and cognitive activity.

14Confirming the statements above, it is important to mention that research into corpus-based translator training has grown considerably in recent years as shown by a variety of publications ranging from practical guides on how to use corpora in the language for specific purposes (LSP) classroom (e.g. Bowker and Pearson 2002) to scholarly volumes which examine and illustrate the use of corpora for pedagogic purposes (e.g. Botley et al. eds. 2000; Hatim 2001; Granger et al. eds. 2003; Zanettin et al. eds. 2003; Laviosa 2008).

15In these studies, students engage with corpus data in a systematic way using a methodology largely based on the principles of data-driven learning (DDL) put forward by Johns (1991a, 1991b) in foreign language teaching. The DDL is an inductive approach, in which students learn bottom-up from data observation. Johns (1991a) advocates the use of microcomputers as informant in language education; the task of the language-learner is, essentially, a “research worker” whose learning needs to be “driven” by access to linguistic data in order to recover the rules from the examples and “discover” the foreign language; and the task of the language teacher (as a director and coordinator of student-initiated research) is to provide a context in which the learner can develop “strategies for discovery” through which he or she can “learn how to learn”. The use of a computer programme (tool: concordancer) can help the learner to develop the ability to see regularities and patterning in the target language and to form generalisations to account for that patterning.

16In this learner-centred context, corpus design and compilation, as well as processing tools and procedural steps—where students identify problem areas arising from Translation Practice, suggest hypotheses, and then test them together with their tutor—are very similar to those employed in corpus-based descriptive research, although the aims are specifically pedagogic.

17In translation pedagogy, small, specialised corpora are designed and used not only as resources for retrieving translation equivalents, particularly in terminological work, but also as repositories of data for improving students’ understanding of actual translational regularities. Moreover, translator trainers working in the field of LSP studies have drawn on the insights of descriptive research into translation universals, and carried out experimental studies aimed at testing in the classroom environment the validity of a number of universals, their usefulness in raising student awareness about translation and translating, and their possible use as predictors of higher or lower quality translations (Laviosa 2008).

18While corpus-based translation teaching is largely inspired by principles consistent with the DDL approach developed by Johns (1991a, 1991b) in foreign language education, descriptive corpus-based translation studies (Baker 1996; Laviosa 2008) adopt a methodology which is largely compatible with Johns’s procedures in corpus for discovering regularities and patternings in the foreign language.

19The methodology employed by Baker (1996) allows a process of gradual discovery of how translators use the target language as well as the identification of facts about the nature of translation, through the examination of word lists, keyword lists and concordance lines. This procedure may be used in a complementary way with Aubert’s descriptive methodology (2006), which requires the student and researcher alike to progress from empirical data to the identification and classification of cultural markers present in the corpus as well as the identification of the options adopted by the translators. What follows is an illustration of how these three methods of enquiry have been integrated in a corpus-based methodology for the student-centred, professionally-oriented translation classroom.

4. Case study

20The corpus-based learning method which we are going to illustrate was implemented during the teaching of the annual subject Translation Practice into English IV of the senior year of the undergraduate course in Translation at a Brazilian State university. As cultural elements referring to specific realities of the source language universe frequently offer challenges to the translators, one of the topics addressed by the discipline was the translation of cultural markers. In order to identify and classify the occurrences of cultural terms in the whole corpus, the procedure of examining lists of—word, statistics; —concordances, collocates clusters; —and keyness results, constitutes a useful resource for the student and researcher. Also, in the light of investigations of the typical characteristics of the language of translation (Baker 1996; Scott 1998; Kenny 2001), another topic addressed was the aspects which contribute to the translated text’s simplification, explicitation and normalisation. The learning objectives were: a) to become familiar with corpora as “one of the computer-aided translation tools and resources available to the professional translator” (Laviosa 2008, p. 3), b) to identify, classify and analyze cultural markers in the source texts, c) to retrieve equivalents or solutions adopted to render cultural markers in the target texts, and d) to discover tendencies related to aspects of simplification, explicitation and normalisation in the translated texts.

21The corpora consisted of: a) an aligned parallel corpus of study made up of the source text Viva o Povo Brasileiro, written by Ubaldo Ribeiro, and its translation An Invincible Memory; and b) a parallel corpus for comparison made up of ten contemporary Brazilian novels and their respective translations into English. In both corpora, the source texts have a common theme which deals with the strong presence of popular culture and highlights manifestations of Afro-Brazilian religions, feasts, habits, legends, as well as a wide range of culturally marked terms, and fragments of Afro-Brazilian language. Their respective target texts describe the kind of world that the translators have recreated. It is a highly vivid and passionate world within its own boundaries, and the English reader is invited to watch it, but perhaps from a distance which reflects marked dissimilarities between source and target cultures. The study uses the WordSmith Tools programme to permit rapid access to surface features of the corpus of study as well as the whole texts of the larger corpus for comparison. In addition, the BNC is used to compare the linguistic patterns characterising translational English with the patterns of original English.

4.1. Stage I

22The first stage in corpus analysis consisted in identifying and classifying the most frequent and significant cultural markers contained in the source text sub-corpus of study, using the WordList and Keyword facility provided by WordSmith Tools, and then examining the English translations using the Concord tool. This enables to focus on cultural terms and exclude other words in the corpus.

Table 1

Cultural domains	Source texts cultural markers	Frequency	Keyness	Target texts cultural markers
Social	negro	118	302.1	black, negro, slave; preto, nego; João Benigno, Feliciano; second coachman, cuddy man
Social	caboco	91	1,111.60	Caboco
Ecological	baiacu	71	820.3	puffer fish
Material	cachaça	30	132.4	cachaça; booze, rum, sugarcane rum, firewater, liquor, sugarcane liquor; omission
Ideological	Iansã	11	91.7	Yansan

Five most frequent and representative cultural markers per domain in the corpus of study

24These cultural markers are registered in Brazilian-Portuguese dictionaries, under the classification of regionalism, folklore or Brasileirismos. The analysis of aligned concordance lines enable the students to infer that the translator is often confronted by the complexities of different social organisations, classes, origins, or practices. To investigate the collocational profile of cultural markers and their respective translations, we used the View button function by clicking the option <grow> to increase the height of concordance lines and show their respective co‑texts.

[VPB] Se NEGO Leléu trabalha? Mas como trabalha o NEGO Leléu! NEGO Leléu ficou forro por testamento de um português de Salinas da Margarida, não quiseram libertar, olhavam para o papel e liam mentiras que não estavam escritas.

[IM] Does BLACK Leléu work? But how he works, that Leléu! BLACK Leléu was freed by the will of a Portuguese from Salinas da Margarida, but was not let go; they would look at the paper and read lies that were not written on.

[VPB] Através dele mesmo, os escravos, PRETOS rudes e praticamente irracionais, encontravam no serviço humilde o caminho da salvação cristã que do contrário nunca lhes seria aberto, faziam suas tarefas e recebiam comida, agasalho, teto e remédios, mais do que a maioria deles merecia, pelo muito de dissabores e cuidados que infligiam a seus donos e pela ingratidão embrutecida, natural em NEGROS e gentios igualmente.

[IM] He himself made it possible for SLAVES, uncouth, practically irrational NEGROES, to find in their humble service the way to Christian salvation, which otherwise would never open itself for them, and to do their chores, receiving in return food, clothes, shelter, and medicines, more than most of them deserved, considering the worries and vexations they inflicted upon on their masters, and their brutish ingratitude, equally inborn in both BLACKS and natives.

[VPB] E mais a espionagem feita pelos NEGROS E NEGRAS da casa do Bângala,

[IM] and also the espionage carried by the SLAVES, in their house in the Bângala district,

[VPB] — Tu, NEGRO ORDINÁRIO, tu eu acerto hoje, hoje eu te acerto.

[IM] You, you STINKING NEGRO, I’ll fix you today, I’ll fix you.

[VPB] Era uma vez, disse, um NEGRO CATIVO FUMBAMBENTO DE CAL que fez para mais de vinte filhos, porém não conhecendo nenhum, que todos levaram embora logo cedo.

[IM] One upon a time, he said, there was a LIME-SOOTED CAPTIVE BLACK MAN who made over twenty children, more than twenty, although he did not get to know any of them, because they took them all away soon after they were born.

26In the quoted novel, the recurring cultural marker caboco/’caboco’ is impregnated with different meanings, some of them going back to the time of settlement when referring to a half-breed man—born from a Brazilian Indian mother and a white father). Other meanings were adapted to its new contexts, such as caboco meaning a “rustic” or an uncouth man. For this reason, this cultural marker is ambiguous even for the students as source language readers. The following are examples of its high frequent collocates and their translations by using loan and transposition: língua de caboco/“caboco language” and cabeça do caboco Capiroba/“caboco Capiroba’s head” in:

[VPB] Assim, não se pode alegar que os padres só obtiveram êxitos, mas conseguiram bastante de útil e proveitoso, apesar de tudo isso haver piorado os sofrimentos da CABEÇA DO CABOCO CAPIROBA.

[IM] Thus it cannot be alleged that the priests were successful in every respect, but they accomplished much that was useful and advantageous, although it worsened the suffering of CABOCO CAPIROBA’S HEAD.

[VPB] gritou Inácia, levantando-se e falando LÍNGUA DE CABOCO muito perto do rosto dele, que curvou a cabeça para trás.

[IM] Inácia shouted, rising and speaking CABOCO LANGUAGE very close to his face, and he moved his head back.

27With regard to the ecological domain, the cultural marker baiacu was selected for analysis. Through the concordance lines, the students could identify that, in the novel, it designates either a type of fish when referring to the ecological domain, or a toponym when related to the material domain. When referring to fish, it is translated through explicitation as “puffer fish”, implicitation by using the pronoun “it”, or a hyperonym “fish”. Regarding the collocate escaldado de baiacu, the translator chose explicitation with adaptation “puffer fish stew”, as shown below:

[VPB] — Mas que lindo ESCALDADO DE BAIACU borbulha no caçarolão de barro, minha gente!

[IM] Folks, what a beautiful PUFFER FISH STEW is a bubbling in the big clay pan!

28Concerning the occurrences of the toponym, the translation options were, respectively, loan or explicitation with transposition followed by an addition of the name of a more well-known location:

[VPB] Dafé lembrou que, se Vô Leléu estivesse no BAIACU e não na Bahia resolvendo negócios,

[IM] Dafé remembered that if Grandpa Leléu were in BAIACU instead of on a business trip to Bahia,

[VPB] ARRAIAL DO BAIACU, 12 de maio de 1841.

[IM] PUFFER FISH VILLAGE, ITAPARICA, May 12, 1841.

29In relation to the material domain, the marker cachaça is rendered by different translation modalities, such as loan transcribed in italics: cachaça; modulation: “booze”; adaptation: “rum”, “sugarcane rum”, “firewater”, “liquor”, “sugarcane liquor”, and omission. These several options revealed some difficulties faced by the translator due to the density of this cultural marker while in the process of rendering its semantic load into the translated text. In the examination of the concordance line of high frequent collocates, the translation of copinho de cachaça shows options for literal translation with adaptation in: “glass of firewater”, “shot glass of sugarcane liquor”, and “shot glass of sugarcane rum”; also, the same kind of translation was adopted for quartinha de cachaça in: “jug of rum”, as shown in the four examples below:

[VPB] Não ia ficar ali no bar, sentado com o terceiro COPINHO DE CACHAÇA, esperando o desfile escolar.

[IM] He wasn’t going to stay in the bar, sitting with his third GLASS OF FIREWATER, waiting for the school parade.

[VPB] — Eu sei, disse Stalin José, levantando-se para pegar um COPINHO DE CACHAÇA no bar.

[IM] “I know,” Stalin José said, rising to get a SHOT GLASS OF SUGARCANE LIQUOR in the bar.

[VPB] Segurando o quarto COPINHO DE CACHAÇA, Stalin José caminhou até a esquina da Rua Direita.

[IM] Holding his fourth SHOT GLASS OF SUGARCANE RUM, Stalin José walked to the corner of Straight Street.

[VPB] — Não, é que ela bebeu quase uma QUARTINHA DE CACHAÇA, a julgar pelo resto que ainda ficou na encruzilhada.

[IM] No, it’s just that she’s drunk nearly a whole JUG OF RUM, to judge from what she left at the crossing.

30Concerning the ideological domain, the mixture of different religions in Brazil usually poses a challenge for any translator of texts containing substantial amount of cultural linguistic markers, for which, probably, there are no satisfactory corresponding terms in the target language. Through word lists and concordance lines, the students could observe that the names of spiritual entities in Afro-Brazilian syncretism are translated by some calques of the source language phonemes and graphemes; for instance, Iansã, orixás, Oxóssi, Xangô, Oxalá and Exu are respectively translated by: “Yansan” , “orishahs”, “Oshosse”, “Shango”, “Oshallah” and “Eshoo”. Besides the calque in the example below, the invocatory expressions is rendered by literal translation and transposition:

[VPB] — Ê-parrê, IANSÃ, SENHORA DOS VENTOS E DAS TEMPESTADES, rainha dos espíritos, valente e ousada como os tufões, de bravura irresistível, eu te saúdo!

[IM, ] Eppa-heh, YANSAN, MISTRESS OF WINDS AND STORMS, queen of spirits, plucky and bold like a typhoon, of irresistible courage, I salute you!

4.2. Stage II

31The second stage consisted in examining the occurrences of the five most frequent and significant cultural markers per domain in the parallel corpus of study in relation to the most frequent occurrences of cultural markers present in the larger parallel corpus for comparison. In an analogous way, the five most frequent cultural markers in the parallel corpus were compared with the reference corpus BNC.

32The same procedure was repeated for the parallel corpus for comparison, and the distribution according to cultural domains was as follows:

Table 2

Cultural domains	Source texts cultural markers	Target texts cultural markers
Social	negro	black (frequency: 362), slave (53), negro (37)
Material	cachaça	cachaça (6), rum (82), liquor (59), booze (2)
Ideological	Iansã	Yansan (71)
Social	caboco	caboclo (9)
Ecological	baiacu	-

More frequent cultural markers per domain in the corpus for comparison in relation to the translation sub-corpus of study

33Out of 555,055 running words of the translated texts of the sub-corpus for comparison, it was possible to retrieve a high occurrence of cultural markers which had appeared in the corpus of study. For the social domain, the term negro shows similar translations of “black”, “slave” and “negro”; however, their respective use shows either a positive or a negative semantic prosody. The following are examples of the three cultural markers extracted from the sub-corpus for comparison:

• and he’d sent a vial of perfume so the Princess of Aiocá (that’s what the BLACKS call Iemanjá) would always have scented hair.

• He’d matured into a calm and cordial BLACK MAN, conserving his simple ways and his proud bearing, his friendly charac.

• I had to back up and then I raised my foot and kicked the BLACK’s BUTT with the tip of my boot and pushed him on top of the others.

• There she combs her hair (beautiful SLAVE GIRLS come with combs of silver and ivory) hears the prayers of the women of the sea.

• No longer treated as a HOUSEHOLD SLAVE, Anastácia began living the life of a respected employee.

• in Ribeirópolis in a couple of minutes, runaway SLAVE, son of a mare, son of a cow, son of a jackass, yellow skunk, bastard, bastard, bastard, bastard!

• When the NEGRO had invited the girl to sit down, he had not done this as a mere formality, or speaking empty words. It was a definite offer; she could choose the place she liked best.

• It was said in the marketplace and vicinity that NEGRO MASSU did not know his own strength.

• but even so he gave as good as he got, and he got plenty. “Give that impudent NEGRO a beating,” Chico Pinóia ordered.

34Another cultural term distributed in the social domain is caboco, which is translated by “caboclo” in the corpus for comparison, but it presents a different meaning usually designating an entity of the Afro-Brazilian religion, as in:

• the Temple of Portão she held in her arms Mãe Mirinha, who had incorporated the CABOCLO PEDRA PRETA, the mixed-blood group of the Black Rock.

• the night in them, dancing, Jesuíno Crazy Cock, now a god, a divinity of the CABOCLO CANDOMBLÉ, a minor god of the people of Bahia.

• a votary who as yet had no special divinity, an eyeful of a mulatta—a new CABOCLO DEITY unknown up to that time.

35For the material domain, the term cachaça presents similar translations of loan transcribed in italics: cachaça; adaptations: “rum”, “liquor”; and modulations: “booze”. Different options in the corpus for comparison are registered with “drink(s, ing)”, “thirst”, “glass”, “white rum”, “cane liquor”. The use of several solutions adopted for certain cultural markers by different translators reveal the increased degree of difficulty when rendering them into the target language and culture. The following examples with the three cultural markers were retrieved from the translated texts of the corpus for comparison:

• circle of luminous dots bloomed in the pitch of night, illuminating bottles of CACHAÇA, dead chickens, and piles of profane medallions.

• Anyhow, whoever eats a piece of jack fruit and drinks any kind of hard LIQUOR on top of it, his skin breaks out all over.

• The conversation took on a certain animation, the empty jug of RUM was replaced by another, and finally Corporal Martim could not resist.

• the owner of the Beacon of the Stars, was going back and forth with a bottle of BOOZE in his hand, keeping count of the number of glasses drunk.

36Regarding the ideological domain, the cultural marker Iansã is translated by several occurrences of the calque “Yansan”, and sometimes accompanied with the name of a Catholic saint or a word used in Afro-Brazilian syncretism, as shown in the examples below:

• Rubem Valentim took the tools and WEAPONS OF YANSAN and broke them down and put them back together again, in his double role

• Before the lights came on in their lampposts, SAINT BARBARA YANSAN had disappeared into the midst of her people. The Press Conference

• greeted Dona Canô, small, withered, and fragile, a jade saint. A DAUGHTER OF YANSAN, she galvanized herself into an insolent agitator, the head of a revolt,

37On the other hand, no occurrences were registered for the term “puffer fish”, related to the ecological domain, nor for “firewater” and “sugarcane”, referring to the material domain in the corpus for comparison.

38With regard to the analysis of the most frequent markers per domain in the translated text sub-corpus of study vis-à-vis similar occurrences in the BNC, the distribution according to cultural domains was as follows:

Table 3

Cultural domains	Cultural markers in the English source texts of BNC	Frequency
Social	black	25,180
Social	slave	1,025
Social	negro	159
Material	liquor	565
Material	rum	446
Material	booze	312
Material	sugarcane	17
Material	firewater	5
Ecological	puffer fish	9
Social	caboco, caboclo	0
Material	cachaça	0
Ideological	Iansã, Yansan	0

More frequent cultural markers per domain in the reference corpus BNC in relation to the translation sub-corpus of study

39Out of 103,785,008 running words of the English originally written texts represented in the BNC, for the social domain the terms “black”, “slave” and “negro” show both positive or negative semantic prosody, depending on the context. This permitted to discover an equivalence of use actually realised for these cultural markers in the translated texts of the corpus for comparison. However, there are dissimilarities in the use of these three cultural markers in relation to the translation sub-corpus of study, in which for less marked collocations generally the word “black” is used, and for collocations referring to more aggressive contexts the marker “negro” tends to appear with higher frequency. The following are examples with the three cultural markers retrieved from the BNC:

• The reasons why so many seek futures in sports, and in rather specific countless other BLACK YOUTHS enthusiastically chasing sporting objectives and, perhaps later, organizing their life’s ambitions around their sporting discipline.

• negative images that are attributed to BLACK PEOPLE and do not feel that and support drug addiction, while BLACK YOUTH is mortally jeopardised by

• turn defending their SLAVES against slave-hunting raids from neighbouring kingdoms

• case of a seller of a FEMALE SLAVE who knowingly allows the buyer

• dreams and aspirations not only of NEGRO PEOPLE but of all Americans

• contesting racist representations of the NEGRO’S SENSUALITY; Barbarous Cruelty inflicted on a NEGRO, from John Gabriel Stedman

40For the material domain, the cultural markers “liquor”, “rum” and “booze” appear with high frequency, followed by “sugarcane” and “firewater”. However, the term “cachaça” did not register any occurrence. The following are examples with the cultural markers retrieved from the BNC:

• to receive a load of contraband LIQUOR this very evening, at dusk

• As I handed him a glass of RUM, I noticed that his eyes we

• The prisoners, drunk on home-made BOOZE, barricaded themselves in a dorm

• made into aguardente (SUGARCANE RUM) but this was said

• punk accepts a glass or two of FIREWATER and is led to the sack

41For the ecological domain, the cultural marker ‘puffer fish’ presents either positive or negative connotations, depending on the contexts of the BNC texts, as in:

• This sharp-nosed PUFFER FISH (left) has a real eye, on the right

• natural metabolisms. The PUFFER FISH is found in warmer parts

• However some fish, such as the PUFFER FISH, produce poisons as a

• of people who are poisoned by PUFFER FISH die, usually within

42On the other hand, no occurrences were registered for the term “caboco/caboclo”, related to the social domain, nor for “Iansã/Yansan”, related to the ideological domain in the BNC reference corpus.

43This procedure was used to try to establish a parameter with the linguistic patterns characterising translational English with the patterns of original English. According to Table 3, it was possible to retrieve representative occurrences of six terms (“black”, “slave”, “negro”; “liquor”, “booze”, “rum”) related to two cultural domains, namely social and material domains respectively, in relation to the translation sub-corpus of study.

44On the other hand, from Table 3 we may also infer that some of the cultural elements in the corpus of study refer to specific realities of Brazilian-Portuguese universe and, for this reason, cultural markers as “caboco”, “puffer fish” and “Yansan” were not perceptible in the original English discursive universe, probably due to the lack of a situation which would highlight the differentiation between both cultures in the texts represented in the BNC corpus.

4.3. Stage III

45In the final stage of the corpus analysis, it was possible to identify some of the general tendencies of simplification, explicitation, and normalisation (Baker 1996). Again, due to classroom time limitation, only a few examples could be selected by the students for closer examination.

46Some of the patterns analyzed in the parallel corpus of study may be identified as simplification due to change in punctuation (Baker 1996; Laviosa 1996). The source text presents long paragraphs, generally with long sentences, and the use of commas, semicolons and dashes. This punctuation contributes to the flow of the original narrative development. Through expanded aligned concordance lines, the students could observe that the translator tries to follow the source text punctuation by not breaking paragraphs. However, the translated novel also presents shorter sentences mainly in the dialogues, by changing the comma into a full stop or a semicolon, probably in an attempt to ease the processing load for the target reader. In the fragment of the translated text below, the students identified an alteration from a weaker to a stronger punctuation mark, besides the regular use of quotation marks in the dialogues translated into English:

[VPB] — Pois é — pensou Amleto, deixando à varanda para ir tomar café —, a verdade é que estou em paz com minha consciência, nunca fiz mal a ninguém, sou um homem prestante.

[IM] “That’s right,” Amleto thought, leaving the porch to go have his break-fast. “The truth is I am at peace with my conscience. I never did anyone any harm; I am a worthy man.”

47Other patterns identified in the corpus of study may be considered as explicitation, since the information inferred from the source novel context become explicit through the use of a supplementary explanatory phrase (“more than twenty”), and the spelling out of an implicitation (“after they were born”) to probably improve the readability of the target text:

[VPB] Era uma vez, disse, um negro cativo fumbambento de cal que fez para mais de vinte filhos, porém não conhecendo nenhum, que todos levaram embora logo cedo.

[IM] One upon a time, he said, there was a lime-sooted captive black man who made over twenty children, MORE THAN TWENTY, although he did not get to know any of them, because they took them all away soon AFTER THEY WERE BORN.

48Another occurrence of explicitation was observed in the target text with the insertion of the appositive “Itaparica”, the name of a more well-known location than the small “Puffer Fish Village”:

[VPB] Arraial do Baiacu, 12 de maio de 1841.

[IM] Puffer Fish Village, ITAPARICA, May 12, 1841.

49Normalisation was identified by the students in terms of a shift from colloquial to formal language. In the source text, there are features of oral language normally viewed as typically spoken by Afro-Brazilians, probably in order to obtain more spontaneous and natural dialogues. In the target text, the translator sometimes normalises dialogues by using more formal language:

[VPB] — PODEXÁ, vá dormir descansado, NÓS CUIDA, PODEXÁ.

[IM] LEAVE IT TO US; you can go to bed without a worry, WE’LL DO EVERYTHING; LEAVE IT TO US.

50On the other hand, through the concordance lines it was also possible to examine a greater number of occurrences of Afro-Brazilian language represented by grapho-phonetic features in the source text being adapted to the possibilities of African American Vernacular English, as shown in the examples below:

[VPB, p. 22] — VOTA — falou o preto, com o mesmo sorriso assustador.

[IM, p. 14] “GOBAH.” The black spoke with the same frightening smile.

[VPB, p. 21] — NGMUNDO.

[IM, p. 14] “FIFFYNIGGA”.

51According to Scott (1998), omission may be seen as another aspect of normalisation. It is used as a resource for manipulating data and avoiding redundancies or explanatory sequences in the source text which seem unnecessary in the target text, as in the example below:

[VPB] — Vota — falou o preto, com o mesmo sorriso assustador. — SIM, VOTA.

[IM] “Gobah.” The black spoke with the same frightening smile.

52Words and expressions in the source text, like Vota at the beginning and Sim, vota at the end of the sentence above, are instrumental in contributing towards constructing the dialogues. Also, a feature of Afro-Brazilian language in the original novel, like: — Vota — / “Gobah.” [back translation: “Come back”], contributes to a feeling of the text having been left as it was spoken (unedited). However, the translator has omitted the repetition of the expression: — Sim, vota, thus diminishing the effect.

53On the basis of the relationships unveiled between the target text and its source text, it was possible to infer some tendencies on the part of the translator who consciously or unconsciously seems to use strategies which can be identified as traces of simplification, explicitation, and normalisation in his rendering of the translated novel in the English language and culture. These tendencies also provide evidence for the distinctive nature of the language of translation, and are consistent with Baker’s (1996, 2000, 2004), Scott’s (1998) and Kenny’s (2001) studies. Moreover, the intimate relationship between language and culture underlying the concrete way in which the translation options of cultural markers were identified by the use of a parallel corpus as a resource for translation teaching/learning appears to be consistent with Laviosa’s (2008, 2009) proposals. However, given the limited size of the corpus of study analyzed, these initial generalisations need further discussion and complementation by an analysis of a larger corpus of study.

54At the end of the three stages in corpus analysis, the students, together with the tutor, came to the following conclusions:

• when a cultural marker presents a dense semantic load, the difficulties faced by the translators increase, and several translation options are used (loans, explicitations, modulations, adaptations, additions) in an attempt to render the different connotative meanings along the target texts;

• the translation options in the corpus of study and the larger corpus for comparison show similarities with certain cultural markers, and dissimilarities as regards a greater number of other cultural terms;

• the uses of cultural markers in the corpus of study, the larger corpus for comparison and the BNC appear to be divergently similar;

• a bilingual glossary of cultural makers accompanied with their co-text may offer a useful contribution to students, translators, researchers and teachers working in the area of translation, since in general bilingual dictionaries do not give actual contextualised examples extracted from corpora.

5. By way of conclusion

55The present case study has focused on an authentic problem raised by the students themselves, namely the translation of cultural markers by means of a corpus-based analysis. In the process of identifying equivalents between source and target texts, the students acquired information about the way in which both discourse and cultural markers were laid down in the two languages. In the future, they may also be able to use corpus evidence to translate new texts (Zanettin 1998). By looking for cultural makers, collocational patterns and their equivalents, students were also able to examine similarities and differences across languages and cultures as well as search for vocabulary and phraseology, thus engaging in a meaning creation activity and developing translator skills. In this way, this study sought to demonstrate how a parallel corpus with high occurrence of culture-bound elements can be used to devise a variety of structured and student-centred activities whose aim is to enhance the students’ understanding of source and target texts, and their ability to produce fluent translations.