Subtitling is a translation technique, which, although it has existed since the beginnings of cinema, has broadened its reach from the decade of 1980 onwards. With the popularization of the videocassette, consumers could watch films in their homes whenever they preferred, and with the advent of cable television in the 1990s, subtitling became increasingly present in our daily lives (NUNES, 2012:20-21).

Although a genre of translation, subtitling has rigid technical particularities, which restrict the translator's decisions. Thus, not only must the translator have ample linguistic and cultural knowledge of the two or more languages with which they work, but they also need to master its rigid/strict technical aspects. Such aspects are imperative so that the subtitles in a foreign language allow the viewers to follow the plot without compromising their understanding. Hence, subtitles present the lines in a "condensed" mode, simultaneously respecting the speed of the speech, the message of the plot, and the viewer's minimum reading time. These aspects may vary from country to country, from subtitling company to subtitling company, from media to media, etc.

In general, typical technical requirements of subtitling include (a) the minimum and maximum duration of a subtitle on the screen, (b) the minimum interval time between subtitles, (c) the speed of the character's speech, and (d) the viewer's average reading time.

Besides such technical requirements, there are matters of style and formatting, which the translator must consider during the subtitling process. We can mention the chosen linguistic register, the privileged syntax, the type of font and its color, punctuation, the positioning of subtitles on the screen, the breaking point of the subtitling lines, among others. Many researchers of Translation Studies have already discussed these matters, including Díaz Cintas & Remael (2007), Díaz Cintas (2001 and 2003), Ivarsoon & Carroll (1998), Karamitroglou (1997 and 2000), Gambier (2003), and Gottlieb (1992, 1994 and 2005).

In the Brazilian context, Elaine Alves Trindade Nunes (2012:40-41) states that it is possible to assign the amount of 14 characters per second (considering spaces and punctuation) to the average reader. In Brazil, translators of subtitles use 32 characters per subtitle in shows broadcasted on television and 42 characters in shows distributed on DVD.

According to Nunes (2012:40-41), the norms of style and formatting follow what the contracting companies determine (in her research, cable television networks). Such stylistic determinations also restrict the translator's decisions, since they tend to homogenize registers, idiolects, and variants. Besides, the linguistic registers suffered deep standardization in their translation into subtitles until 2010 with the deletion of regionalisms and slangs; the softening of profanity and foul language; and the sole use of standard written language to express the uniqueness of spoken language since it has its own grammar rules and usages (NUNES, 2012:48).

Based on Nunes’ observations and her experience in subtitling with Brazilian and foreign companies (both American and European), she developed a brief compilation of techniques where she summarizes practices demanded by companies in Brazil and abroad. For this purpose, she consulted the following documents: HBO Subtitling Manual (2007) and Superstation Subtitling Manual (2001), both from Brazil; Kitchen TV Channel Subtitling Manual (2009), from the USA; and Undertext Subtitling Manual (2009), from Sweden³.

For subtitles in English, we found similar values in materials from different companies. Netflix’s support page suggests up to 42 characters per line of subtitles (What is the maximum, n.d.). Captiz⁴, a website that promotes freelance subtitling services trades, suggests less than 20 characters per second, therefore less than 37 characters per line (Guidelines for video subtitling, n.d.). Permondo⁵, an NGO that does translations voluntarily, suggests no more than 35 characters per subtitle line (Introduction to subtitling, n.d.).

These requirements make subtitle translation a complex task, especially when the translator faces linguistic elements specific to a culture and unknown in the reception culture of the work. In other contexts and stemming from distinct theoretical perspectives, these elements obtained attention from many scholars. For instance, Nida (1945, cultural domains), Fiódorov (1953, Soviet School and Comparative Studies), Vinay and Darbelnet (1958, Structuralism and Stylistics), Otto Kade (1964, The Leipzig School and its studies concerning the concept of Realia), Vázquez-Ayora (1977, Linguistic Studies) and Oksaar (1988, Cultural Theory), whose texts are well known. As is common knowledge, different points of view elicit designations and concepts, which, although similar, cannot be reduced to one another, especially when considering their theoretical basis. As examples, we can cite the works of Aixelá (1996), Nord (1997), Mayoral & Muñoz (1997), Azenha (1999), Martínez (2006), Aubert (2006a), Nadal (2009) and House (2018). All of them are coherent with the theoretical perspectives that inspired them and diverse in their particularities.⁶

Although it is possible to say that our research dialogues with some of these works, especially with Aubert (2006a), we adopted the perspective developed by Reichmann & Zavaglia (2014:52) and Zavaglia (2018)⁷ of cultural marks and markers.

Over time, cultural markers have never impeded translations, as there is a significant variety of translation strategies available to the translator, whose creativity can multiply them. In the case of audiovisual materials, the translator also relies on extra-linguistic elements, such as the soundtrack, speech intonation, image, all of which complement the text and promote understanding. In deference to the unique nature of subtitling, we propose a typology of translation of cultural markers that takes the subtitling particularities into account, namely its multimodality. The typology contrasts to various works that have already discussed the cultural matter (and others topics on this domain), such as the works from Perego (2003), Bruti (2009), Hawkins (2010), Guillot (2010), Zojer (2011), Pedersen (2011) and Oliveira & Kilian (2016).

In this article, we initially present the background of the film O auto da Compadecida (A dog's will, Arraes, 2000). Secondly, we discuss the concept of cultural markers that pervade our approach to the topic and the translation modalities based on Aubert (2006a, 2006b). Thirdly, we introduce a typology of subtitles according to the extra-linguistic elements present in a scene. Finally, we combine the typology of subtitles with the translation modalities, providing examples taken from the film. Thus, the analysis considers the text of each subtitle, the combination of textual meanings of each cultural marker with the aid provided by the audiovisual elements, and the technical restrictions of subtitling.

We chose the Brazilian film O auto da Compadecida (A dog’s will, 2000), directed by Guel Arraes and adapted from the homonymous play by Ariano Suassuna, as the material of our analysis. The motivation for this choice was the significant number of cultural markers from the 1930s, especially from the states of Pernambuco and Paraíba, in the Northeastern region of Brazil. It is relevant to mention that many of these markers are still valid in the Northeastern way of speaking and other regions of the country.

O auto da Compadecida was part of the Brazilian film movement known as Cinema da Retomada (literally, Cinema of the Retaking): after two and a half decades of the decadence of national cinematographic production, the public returned to the cinema, albeit timidly, to appreciate and honor Brazilian films. This change resulted from the incentive laws promulgated in the country beginning in the early 1990s. O auto da Compadecida received national and international awards.

Among other reasons that favored our choice are: the careful linguistic acuity of the script, the aesthetic and historical meticulousness of the work, and the importance of the film due to the considerable number of viewers it took back to movie theaters, having been screened in a miniseries format a year earlier by the television network Rede Globo.

Such factors express the appeal that the work exerted - and still exerts - in the Brazilian collective mind. It is currently considered one of the masterpieces of Brazilian national cinema.

With its language full of humor and long, rapid dialogues, the film tells the story of two poor, illiterate sertanejos (citizens from arid areas of Northeastern Brazil) living in the town of Taperoá, in the countryside of the state of Paraíba.

The historically underprivileged region is part of the Northeastern sertão (remote dry areas) and faced the political dominion of oligarchies of wealthy landowner families, the so-called coronelismo (or rule of the “colonels” - patriarchs of wealthy families). In response to this political domain, to the social needs, and, frequently, personal disagreements, the Cangaço movement emerged. Considered social banditry, which was in force in the mid-nineteenth and early twentieth centuries, it left marks in the culture and the Brazilian language variant.

The main characters of the story, João Grilo, and Chicó, are poor, hungry, uneducated, without family or possessions. To survive, they try to get any work they can in town: they do personal favors and take on side jobs. Chicó is naïve, has a good heart, and is extremely fearful, whereas João Grilo is clever and uses his cunning nature to deceive his fellow citizens for his benefit.

Along with the story, the two interact with several characters, who personify different social groups in a caricatured way: the miser baker Eurico and his unfaithful, lusty wife Dora represent unjust bosses. Father João and the town bishop personify religious hypocrisy and the dynamics of submission to the rich and negligence to the poor. Corporal 70 represents low-level authorities. Vicentão is the town bully, who disputes against Chicó and Corporal 70 for the love of Rosinha, Major Antônio Moraes' pure but insightful daughter. Major Antônio Moraes represents the political authority of the town, and Severino de Aracaju is the cangaceiro (a regional bandit) who wants to take revenge on the negligent population of Taperoá.

In the second part of the film, Severino dies by one of João Grilo’s tricks, and his henchmen kill almost all characters who then wait for the final judgment. In the end, the characters face their trials as they follow the argument between Jesus Christ, Our Lady Mary, referred to as the Compadecida (the Compassionate), and the Devil.

To define cultural markers, we found in Anthropology the perspective to support the definition of culture that pervades the concepts of cultural markers and cultural marks we use in the research. Anchored in Laraia (1999), we came to the understanding that culture encompasses all material and immaterial manifestations created by humans, which guide their lives and behavior as members of a group, hence their habits, customs, political institutions, social organizations, economics, beliefs, tools, among others. These human creations coexist in the development of a group, producing a unique collective experience in unique historical, political, social, and economic contexts.

According to Reichmann and Zavaglia (2014:45-46), this specific worldview - although other groups might share certain aspects of it - generates an abstract collective way of thinking inherent to the reality of the group that shares it. This abstract collective way of thinking represents what we call cultural marks, and from these cultural marks derive concrete language uses, known as cultural markers. Therefore, we should only consider a lexical item, a phoneme, or even a textual genre as a cultural marker if, in their use, we can observe the presence of cultural marks.

To illustrate the concept, we mention Reichmann & Zavaglia's research (2014) on Brazilian school documents translated into French and German. These documents are a concrete manifestation of a collective thought shared by Brazilian society regarding Education. It means that the abstract cultural organization that generates these documents and their uses (i.e. selected information, the location of each piece of information on the paper, the necessary signatures, the form of validation, the governmental bodies cited on the document, the social situations that require such documents to be printed and delivered to a person, for example) reflect the concept of Education specific to the Brazilian society.

Many aspects can be - and are - shared by other socio-cultural groups since they influence each other. However, their specific linguistic set as a document (a Brazilian diploma or school transcript, for example) is unique to the Brazilian culture. Thus, the conception of Education shared among Brazilian people constitutes a cultural mark, while school documents and their specific contents are cultural markers.

Although cultural markers can manifest themselves in any dimension of language - from phoneme to textual genres - in this article, we only focused on lexical items. The analysis initially used Corpus Linguistics tools (from the software AntConc⁸, version 3.4.4, Anthony, 2016), to match the film script in Portuguese and their respective subtitles in English.

As a second step, we used the software to compare two corpora (a general reference corpus with 2,243,823 words extracted from Lácio-Ref⁹ and the study corpus, the film script containing 16,701 words) and generate a frequency list from the corpus in Portuguese. From the lexical items on this wordlist, we observed each use in context employing AntConc's concordance tool to identify candidates for cultural markers.

During this process, we observed that we could group the candidates thematically. Therefore, we divided them into the following categories: (a) the characters, (b) Northeastern religiousness, (c) toponyms, (d) fauna and flora, (e) cuisine, (f) sertão, (g) cangaço and (h) currencies. Finally, we analyzed each marker in its audiovisual context. Through this analysis, we could observe the lexical items in the subtitles concerning the non-verbal elements of the film and the cultural marks. We ultimately reached 101 cultural markers.

After we reached a list of cultural markers, we classified them according to Aubert’s translation modalities (1999, 2006a, 2006b), inspired by the proposal of Vinay and Darbelnet (1958). We briefly present the modalities below (Aubert, 2006b:60-69) and exemplify with lines taken from the film O auto da Compadecida (A dog´s will, 2000).

To better understand the translation modalities, we provide, for each example, the speech followed by the exact moment it happens in the film, the subtitle available on DVD,¹⁰ and a literal translation made by us.

1. Omission: a given textual segment of the source language is not translated and its reference cannot be rebuilt in the target text. In the example below, the toponym Cabaceiras was omitted in the translation. It is not possible to reconstruct this information whether by text or by extra-linguistic elements available in the film.

2. Mirroring: the translation presents little structural and semantic change concerning the source text. It is divided into:

2.1 Loan: the translation is identical to the translated segment in the source language. In the example below, we notice that the translation of the toponym Glória de Goitá was loaned. That is to say, that it used similar spelling as in Portuguese.

2.2 Calque: the translation is almost loaned from the source language, but it undergoes graphic or morphological changes according to the conventions of the target language. In this example, the name Antônio was translated with spelling accommodations.

3. Literal translation: the translation into the target language is an intercultural synonym of the translated segment of the source language. It is divided into:

3.1 Transcription: the textual segment in translation belongs to the two linguistic systems considered (mathematical numbers or equations) or to none (textual segments in a third language). We did not find examples of transcription in the corpus.

3.2 Word-for-word translation (or WWT): the textual segment of the source language is translated into the target language with the same number of words, in the same syntactic order, with the same grammatical categories. The example below represents an oral poem the character recited to invoke Our Lady (Nossa Senhora, Mãe de Deus de Nazaré) so that she could intercede on his behalf in the final judgment. The second occurrence of the marker was translated by WWT. Due to the long length of the line, the subtitles were divided on-screen as follows:

3.3 Transposition: it occurs when one of the requirements of the word-for-word translation does not hold. Note the cultural marker Padre João (Father João) in the following example:

3.4 Explicitation: it occurs by paraphrases, appositions, parenthetical explanations, footnotes, glossaries, prefaces, or afterwords (see comment below).

4. Equivalence: the source text is translated from an interpretive perspective of the target culture. It is divided into:

4.1 Implicitation: the textual segment of the source language is not translated, but it is possible to reconstruct its reference in the target text (due to the previous mentioning of the cultural marker, for example). In this example, the toponym Recife was previously mentioned in the film, so the viewer can understand that Rosinha travels from Taperoá to Recife.

4.2 Modulation: the translation presents a different point of view from the textual segment of the source language, but the constructed reference is still based on the source. In this example, the cangaceiros used macaco (monkey) to offend the militia that chased them. Although it might have a racial connotation, it refers to the way cangaceiros saw the militia, namely as men who only knew how to follow orders.

4.3 Adaptation: the translation builds a new reference based on elements of the target culture. In this example, contos de réis represent a Brazilian currency used from colonial times until 1942. In English, the translator used the British coin crown.

5. Intersemiotic translation: a translation that implies a non-verbal system.

6. Error: an erroneous translation. It is not possible to justify the changes in perspective. In the example below, Dona Dora, which in Portuguese means Mrs. Dora, was translated to Dona Dora by loan. However, for a non-native viewer, it gives the impression that the character’s name is Dona (spelled with one n). The word dona is an honorific that, in this case, shows courtesy and respect. Her interlocutor would not have addressed her by her first name since he is her employee and considerably poorer than her.

It is possible to observe two or more modalities in the description of the translation of the same textual segment. However, one of them tends to prevail.

Through the collation of the source text and the translated text, the translation modalities allow the observation, on a case-by-case basis, of the strategies the translator used. However, since they essentially deal with written translation, we felt the need to revise and redefine them, adapting them when necessary.

One of the cases that merits highlighting is the modality of explicitation. In a written text, explicitation becomes an interpretative tool for the reader since it provides more information about the text. In the context of films, the cover, the posters, the publicity materials, forums, and sites about television and cinema on the Internet can perform a function analogous to that of explicitation, as a form of audiovisual epitext (Genette, 1987). The subtitle may assume the form of an audiovisual peritext since it explicits - even if in another language - the audio content. In its turn, we may notice it more punctually in the lexicon of subtitles, which occurs peculiarly. Note the word farinha on the following example:

The association of the lexical item farinha with wheat flour is commonplace, as it is abundant in the totality of Brazilian cuisine and countless other countries. However, in this context, the word farinha allows the Brazilian viewer, instantaneously, to assume that the character is referring to manioc flour, a product widely used as one of the bases of Northeastern cuisine and an element of the sertanejo’s subsistence.

The translation was manioc - by modulation and explicitation since the specific item presents another point of view for the reader of the meta-text and, at the same time, makes one of the characteristics of the referent explicit, respectively. It explicits the cultural matter to the viewer of the reception culture, undoing possible associations with any other type of flour had the translator preferred the literality of flour.

Once we collated the source text and its respective translations and analyzed the translation modalities used for each selected cultural marker, we began to observe how these cultural markers related to the extra-textual elements present in the film. Therefore, we considered textual and stylistic aspects, subtitling, and audiovisual techniques (sound and image).

Textually, technically and audio-visually, we considered the following: the modalities used for the translation of each marker, the style of the language, the relationship established between the duration of a subtitle on the screen, the number of characters available, and the moment of the work in which we ascertained the first occurrences of the markers. This initial moment establishes at which point in the film the viewers first confronts themselves with a marker, what scenes came before that moment, and how they relate to the first occurrence. Concerning image and sound, the following were considered: the scenery, the body language of the characters, the intonation of the speech, the soundscape, and the soundtrack.

In the course of these observations, we detected three groups of subtitles, which we named introductory subtitles, verbal subtitles, and subtitles with image and sound support, as defined as follows:

1. Introductory Subtitles: they present for the first time a piece of information or lexical item, which is relevant to the understanding of the plot or the characteristics of the characters and is responsible for shaping the perspective of the viewer on the subject. They do not dispense with the contributions of the text itself or the image and the sound. Therefore, they are also verbal subtitles or subtitles supported by image and sound.

2. Verbal Subtitles: the subtitle's textual support is the most relevant - or the only one available - to understand either the plot or the characteristics of the characters. Audiovisual elements are dispensable for understanding or nonexistent, and the text is the only context available to the viewer. In this case, the technical restrictions of subtitling, such as the time and number of characters available on the screen, are influential, so we may also refer to them as Spatial-Temporal Subtitles.

3. Subtitles with support from the image and sound: considering the technical restrictions to which they are submitted, image and sound offer support for comprehending the meaning of cultural markers. They were divided into:

3.1 Background-Soundscaped Subtitles: in this case, photography, soundtrack, and soundscape are indispensable for understanding the text. Thus, the background plays a crucial role in reconstructing the plot. These are the most frequent types.

3.2 Gesture-Intoned Subtitles: in this case, the gestures of the characters, as well as the intonation of their lines, are indispensable for the comprehension of the text. Thus, the foreground helps comprehend the text.

The three types of subtitles described above - among each, one is subdivided into two groups - aim to consider, in the comparison of the translations and their audiovisual contexts, the highest quantity of extra-textual elements that make up this type of material and how they shape the viewer’s perspective.

We combined each of the types of subtitles, as categorized in the previous section, with Aubert's (2006b) translation modalities to establish a typology of translation of cultural markers in subtitles, as shown in the following table:

Table 1

Subtitle	Introductory Subtitle	Spatial-Temporal Subtitles	Background-Soundscaped Subtitles	Gesture-Intoned Subtitles
Modality
Omission	Omitted	Omitted	Omitted	Omitted
	Introductory	Spatial-Temporal	Background-Soundscaped	Gesture-Intoned
	Cultural Marker	Cultural Marker	Cultural Marker	Cultural Marker
Loan	Loaned	Loaned Spatial-	Loaned	Loaned
	Introductory	Temporal	Background-Soundscaped	Gesture-Intoned
	Cultural Marker	Cultural Marker	Cultural Marker	Cultural Marker
Word for Word (WWT)	WWT	WWT	WWT	WWT
	Introductory	Spatial-Temporal	Background-Soundscaped	Gesture-Intoned
	Cultural Marker	Cultural Marker	Cultural Marker	Cultural Marker
Implicitation	Implied	Implied	Implied	Implied
	Introductory	Spatial-Temporal	Background-Soundscaped	Gesture-Intoned
	Cultural Marker	Cultural Marker	Cultural Marker	Cultural Marker
Error	Erroneous	Erroneous Spatial-	Erroneous	Erroneous
	Introductory	Temporal Cultural	Background-Soundscaped	Gesture-Intoned
	Cultural Marker	Marker	Cultural Marker	Cultural Marker

As may be observed in the table above, the strategies for translating cultural markers, in the context of subtitling, can be of 44 types resulting from the intertwining of subtitle types and modalities.

Regarding the object of the present analysis and the typology above (Table 1), we observed 31 types of translation of cultural markers in subtitles. However, as transposition was abundantly found in combination with other more relevant modalities, they were not cataloged as transposed subtitles, resulting in 27 cataloged types¹². We did not identify nine types: erroneous introductory subtitles, erroneous spatial-temporal subtitles, omitted background-soundscaped subtitles, erroneous background-soundscaped, and loaned, explained, implied, adapted, and erroneous gesture-intoned subtitles. The same occurred with the transcription (four types of translation), which is the modality that designates the translation of lexical items which are part of the two cultural systems, or items that refer to a third culture, common in the subtitles of scientific documentaries (formulas, for example). Such markers did not appear in the research. ¹³

To illustrate some of the types found, we present the examples below.

Omitted Introductory Cultural Marker: in its first relevant occurrence in the film, the marker was omitted. Lacking the support of image or sound, the viewer cannot reconstruct its cultural mark.

In Portuguese, the honorific senhor (mister), accompanied by a proper name (as in Senhor Pereira - Mr. Pereira) or by a position or function (as in Senhor Major - Mr. Major or Senhor Presidente - Mr. President), introduces formal hierarchical treatment and respect. In the Brazillian pronouns system, speakers use senhor in opposition to você (you) or tu (thou used as you in some regions) when addressing a man. This use has become commonplace in the face of economic differences, differences of position, and/or age between interlocutors. Seu is the common corrupted form of senhor.

In other cases, however, seu may acquire a sarcastic or ironic tone to "make the interlocutor notice attention, pretense, pride, or even lack of modesty, usually when addressing a younger speaker (breaking the rule)" (Piccolo, n.d., our translation).

Note the use of the honorific seu in the following example:

Example [1]:

Picture 1 Auto da Compadecida, O. Directed by: Guel Arraes, Production: Globo Filmes. Rio de Janeiro: Sony Pictures do Brasil, 2000, 2 DVDs (00:25:08).

In this scene, Chicó begins a love affair with Dora and - like the coward that he is - asks his lover what to do if Eurico, her husband, surprises them. As Chicó is an employee of Eurico - and poorer than him -, the speech is marked by the use of seu before the name, despite the betrayal, showing respect and submission (cultural mark).

In the subtitle, the omission of the title modifies the relationship between Eurico and Chicó, and the audiovisual elements do not offer aid for its reconstruction. In this case, Chicó is both coward and respectful, but the viewer cannot reconstruct his respect towards Eurico, marked in Portuguese by the use of the pronoun seu before his name, because it was omitted in English (a possible translation in English, by modulation, would be mister).

This omission was likely due to technical restrictions since the translator would have approximately 1.5 seconds and up to 21 characters for Chicó's speech, but the subtitle, including the omission, has 24. However, this is a supposition. Nevertheless, not all cases of Omitted Introductory Cultural Markers have, as a restriction, the number of characters available concerning the character's speaking time, which configures a choice of the translator to omit them¹⁴.

Loaned Introductory Cultural Marker: in its first occurrence, the marker was translated by loan. Sound and image support the understanding of the first occurrence of the cultural marker pirarucu (one of the largest freshwater fish in Brazil, which inhabits the Amazon basin and can reach 3 meters in length) in the film. The scene shows Chicó along with the fish by the river. This subtitle is also a background-soundscaped subtitle.

Example [2]:

Picture 2 Auto da Compadecida, O. Directed by: Guel Arraes, Production: Globo Filmes. Rio de Janeiro: Sony Pictures do Brasil, 2000, 2 DVDs (00:14:35).

Although a coward, Chicó is known in town for telling all sorts of fibs. These fibs are usually absurd stories where he is frequently the protagonist. However, these stories happen in many different parts of the country, which endorses that he is likely lying.

In the scene, Chicó recalls the beautiful words he (had) heard from the priest when his pirarucu died. His friend, suspicious, questions him about owning a fish of this size and referring to it as a pet. Chicó then tells his unbelieving friend that he believed that it was the fish that owned him since, when trying to catch it, he was trapped in the harpoon rope and then dragged by the animal upstream for three days. In the scene, the viewer can have a visual image of the pirarucu fish, although caricatured.

Calqued Gesture-Intoned Cultural Marker: the translation strategy of the marker, with support from gestures and the characters' intonation, is the calque. In the following example, note the name Manuel.

Example [3]:

Picture 3 Auto da Compadecida, O. Directed by: Guel Arraes, Production: Globo Filmes. Rio de Janeiro: Sony Pictures do Brasil, 2000, 2 DVDs (1:12:46).

Manuel, a proper name of Hebrew origin, is very popular in Portugal and, by extension, in Brazil. However, in Brazil, it refers in some situations to joking and stereotyped associations with the Portuguese, much used with a pejorative meaning. In the film, João Grilo's speech addresses this matter in this scene. João Grilo apologizes to Jesus for not wanting to be disrespectful but says that he must comment on the fact that the Devil called him Manuel, in which he assumes the use of a stereotype or lack of respect for being a common name. The calque Emmanuel does not account for this aspect of the word in Portuguese. However, it is possible to infer the connotation by the gestures, the features, and João Grilo's voice, and his tone of disdain. In this manner, the strategy enables the viewer to reconstruct the cultural mark to which the marker refers.

Explained Background-Soundscaped Cultural Marker: the translation becomes accessible through the change of register, which favors the understanding of a rather specific marker added to the images and soundscape. We will present the marker peixeira.

The dictionary entry for peixeira defines it as a "wide-blade knife used for cleaning fish" or a "short, pointed knife used as a weapon". It is still used as a weapon, and its association with crimes, which happen, notably - but not exclusively - in Northeastern areas of the country, contributed to consolidating the link with the region, reinforced by the fact that it was customary to see armed civilians in the villages. This custom may still exist in some areas of the sertão (remote arid regions of Northeastern Brazil), but it is no longer commonplace in large cities.

Example [4]:

Picture 4 Auto da Compadecida, O. Directed by: Guel Arraes, Production: Globo Filmes. Rio de Janeiro: Sony Pictures do Brasil, 2000, 2 DVDs (00:27:31).

In the scene, Eurico knocks on the door and surprises his wife, Dora, who is in the company of her two lovers at home (Vicentão and Chicó). To find a way of not being caught in adultery, she asks Vicentão, the town bully, to pretend to have come to her house to kill Chicó for any reason. She asks Vicentão to pick up his peixeira and go out shouting that he will kill Chicó. Chicó would be hiding in the closet to prove that he was fleeing from the fight.

Peixeira was translated by explicitation - and modulation - into knife. Although briefly, the viewer can see the type of knife and Vicentão walking with it on his waistband, according to the custom.

Implied Background-Soundscaped Cultural Marker: the image or soundscape allows the viewer to understand the cultural mark of a marker that was implied in the subtitle due to previous occurrences or to the context. To exemplify, we will present one of the occurrences of the marker cangaceiro.

The cangaceiro is seen as a type of regional bandit and has been a representative figure of the Northeastern sertão to the present day. At the end of the nineteenth century and the beginning of the twentieth century - the period when the Cangaço, a social banditry movement, emerged -, the stemming image of the cangaceiro appeared.

The cangaceiro wandered in groups, dressed in leather, armed, and were nomadic, because they knew very well the dry region of the sertão. They lived on looting, donations, and the help of coiteiros (owners of smallholdings and farms who provided them with hiding places from the police). They had their laws and fought against social injustices promoted by ineffective governments. In this manner, they often conflicted with the volantes: small state militias hired to arrest and kill them. Some cangaceiros worked for wealthy landowners, who represented the greatest local authorities. Although some groups struggled against injustice, the cangaceiro spread fear and violence through the villages where they passed. In the film, Severino de Aracaju is the chief of the cangaceiros who invade the town of Taperoá. He takes the villagers hostage and demands that his henchman kill them.

Example [5]:

Picture 5 Auto da Compadecida, O. Directed by: Guel Arraes, Production: Globo Filmes. Rio de Janeiro: Sony Pictures do Brasil, 2000, 2 DVDs (1:20:13).

In the scene, the Compadecida - one of the names given to Our Lady, which means the Compassionate, that one who shows compassion towards others - intercedes with Jesus Christ on behalf of Dora and Eurico. The saint recalls that a moment before the cangaceiro from Severino's group shot the couple, the husband forgave his wife for her affairs. Although the cultural marker cangaceiro does not appear in the subtitle, it is possible to see its reference since the cangaceiro shares the scene with Eurico and Dora. Although the cultural marker does not appear in the subtitle, it was implied by the image.

Modulated Background-Soundscaped Cultural Marker: the marker is translated by modulation, although the context or the images allow the viewer to reconstruct, however partially, the mark. In this type of subtitle, we found more occurrences - in quantitative terms - and the most relevant translations.

In the scenes below, we present the marker papo amarelo - a type of gun called yellow neck - whose translation modulates the register; however, the images allow the understanding of the plot.

According to Carvalho (2008, Oct. 29, our translation), papo amarelo is a popular name for the Winchester rifle, model 1873, caliber 44:

Picture 6 Neto, Carlos. F. P. Rifles e Carabinas Winchester “Lever-Action”.

Example [6]:

Picture 7 Auto da Compadecida, O. Directed by: Guel Arraes, Production: Globo Filmes. Rio de Janeiro: Sony Pictures do Brasil, 2000, 2 DVDs (00:55:07).

In the film, Severino de Aracaju and his men attack the town of Taperoá to rob its townspeople. Severino invades the church and announces the robbery. The Bishop replies by stating that he is poor and has no money to give. Not believing the Bishop, Severino declares that he is even poorer than they are since there is no one to defend him in life, but ‘his old and poor papo amarelo’, that is, the old rifle with which he plundered cities and committed murders.

Papo amarelo was translated by modulation - and explicitation - to old gun, which refers to the cangaceiro's custom of always carrying his rifle. The scene shows the type of rifle, very common during the Cangaço. The demonstrative pronoun this helps refer to the weapon displayed by the character.

Adapted Background-Soundscaped Cultural Marker: the translator uses a specific concept of the target culture to favor the understanding. We will present the case of the marker tostão (plural: tostões), currency translated into crown(s).

Although they are not an exclusive facet of Northeastern's culture, we chose coins as cultural markers because they help to palpably observe a significant cultural mark of the economic policy of the time in which the story takes place and because they are meaningful to the plot.

Example [7]:

Picture 8 Auto da Compadecida, O. Directed by: Guel Arraes, Production: Globo Filmes. Rio de Janeiro: Sony Pictures do Brasil, 2000, 2 DVDs (00:02:00).

In the scene, João Grilo and Chicó go to the bakery to apply for a job and negotiate the salary. The context and the introductory subtitle of the marker help the viewer understand that the topic of the conversation is money. The marker tostões was translated by adaptation to the British coin crowns.

Audiovisual translation has been the object of many studies and has gained considerable attention in Translation Studies. Although it offers many challenges for the translator/subtitler due to its technical, stylistic, and linguistic constraints, it becomes, at the same time, prolific data for the researcher.

This article represents our intention to contribute to the analysis of audiovisual translation and shares a part of our findings, considering the translation of Brazilian cultural markers into subtitles in English. Our objective was to observe how the translation retraces the cultural specificity of a lexical item through the subtitle with the aid from the sound, scenario, soundtrack, intonation, and gesture from the characters.

We combined Reichmann and Zavaglia's (2014) definitions of cultural marks and markers with Aubert's translation modalities (2006a, 2006b). From this combination, we developed a typology of translation of cultural markers in subtitles to analyze audiovisual translations, considering different textual and extra-textual elements present in these materials. This typology helped us analyze, reconstruct and understand the translation of cultural markers.

Our typology consists of three categories of subtitles. While the first considers the moment when the plot introduces a cultural marker to the viewer, the second encompasses subtitles where image and sound cannot anchor the textual translation. Finally, the third one deals with subtitles where audiovisual aids complement the textual understanding.

We analyzed the cultural markers of the Brazilian film O auto da Compadecida (A dog's will, Arraes, 2000). With the support of the Corpus Linguistic program AntConc (Anthony, 20016) we extracted a wordlist with 101 cultural markers. We could observe which types of subtitles and which translation modalities were more frequent, and which ones were not adequate to analyze the translation of cultural markers in our research corpus.

From the 44 types of subtitles of the typology we elaborated, we found 31 types. However, we did not catalog four types of transcripted translation due to its abundance and combination with more relevant modalities, resulting in 27 catalogued types. We did not observe nine types of translation and did not find transcripted cultural markers on the corpus. The category we found the fewest variety was the gesture-intoned subtitles.

By considering the non-textual elements of the audiovisual material, we believe that the typology of subtitles we developed added to the revision of the concepts that supported our research promote a broader look at the translation of cultural markers in subtitling. However, it is still necessary to conduct further research with different corpora to test and improve the typology.