Linguistic variation in Italian neorealist cinema: A multimodal analysis of subtitling

Abstract This paper focuses on the standard and non-standard language varieties present in the original dialogue, i.e. the source text (ST), of Miracolo a Milano by Vittorio De Sica and its English subtitles, i.e. the target text (TT). A two-level multimodal analysis was performed on an annotated corpus to identify the different diegetic functions established by the intermodal relations present in the ST and to assess the impact of the adopted translation strategies in preserving, modifying, or eliminating these intermodal relations in the TT. The results of the textual level of analysis show that the non-standard oral variety is used in the TT as the only solution to translate the non-standard varieties of the ST. However, this choice does not convey information about the geographical origins or social background of the characters, and only signals the informal register of the conversation. At the diegetic level, in the TT, the standard variety associated with the wealthy characters and the non-standard oral variety associated with the slum dwellers maintain the intermodal relations and diegetic functions established in the ST. In some cases, however, the non-standard oral variety in the TT eliminates or modifies the diegetic functions originally assumed.


Francesca Raffi is Senior Research Fellow in
English Translation at the University of Macerata (Italy) where she has been teaching English language and translation at undergraduate and postgraduate levels since 2015. She holds a PhD in English for special purposes and audiovisual translation from the University of Naples Federico II and she is Chartered Linguist of the Chartered Institute of Linguists (UK). Since 2016, she has been collaborating (teaching and research) with the Department of Translation Studies, Constantine the Philosopher University in Nitra (Slovakia) and with the Department of Modern Languages and Cultures, University of Liverpool (UK), where she became an Honorary Fellow in 2019. Her research interests are in the field of linguistic (including interlingual and intralingual translation) and sensorial accessibility (including intersemiotic translation) to texts, media, arts, and culture. She has been involved in several projects in Italy and abroad and has given invited lectures and seminars at various Italian and international universities and institutions. Her list of publications is available here: http://docenti.unimc.it/f.raffi#content= publications

PUBLIC INTEREST STATEMENT
Numerous studies in the field of audiovisual translation have focused on the translation of linguistic varieties in films and TV series since the 1990s. However, the research conducted so far is often limited to the analysis of the monomodal and textual dimensions of communication, overlooking the fact that audiovisual products are multimodal by nature. Given the complex semiotic system of films, it is essential to focus not only on the verbal components, but also on the relationships established between the different modes at play. Therefore, the present study aims to identify the intermodal relations established in the film Miracolo a Milano (Vittorio De Sica, Miracolo a Milano (Miracle in Milan), directed by Vittorio De Sica, 1951, Italy) between different modes: spoken and mise-en-scène modes for the original dialogue in Italian; speech, mise-enscène, and subtitle modes for the English translation. The aim is to assess the impact of the adopted translation strategies in preserving, modifying, or even eliminating these intermodal relations in the subtitles.

Introduction
Linguistic variety can be defined as the linguistic features or patterns of spoken discourse used in specific social and cultural contexts and associated with extralinguistic variables. These communication tools can be used in different combinations to convey the identity of speakers and groups. Indeed, within any language, diverse linguistic features serve different functions depending on the meaning that speakers attribute to them in a given social or geographical context, which will also be determined by the communicative situation and the medium used.
In audiovisual products, linguistic varieties strengthen the credibility of character profiles, thus allowing the viewer to complete and enrich their characterisation. In the words of Ramos Pinto (2017: 4), "the creative use of non-standard varieties in audiovisual products functions as a mimetic resource [. . .] for the indirect depiction of the characters, the interpersonal relations established between them and the discursive situations". In particular, subtitling involves a diasemiotic transposition, where there is a shift from the verbal communicative channel (speech) to the written word, which requires an oral text to be reconstructed and condensed into strings of written text. The translation task is further complicated when the translator is confronted with the director's choice to characterise the profile of a character by attributing to them a particular cadence or a specific non-standard variety.
In this context, Italian neorealist films 1 became emblematic in representing social, cultural, and economic distinctions through the resources of language. In fact, dialect is used not only to faithfully represent the pluralistic world of post-war Italy but also to create a binary opposition, positive versus negative, reflected in the contrast between dialect and standard Italian: on the one hand, the positive code represented by lower-class Italians and associated with dialect; on the other, the negative code represented by middle-class characters speaking in standard Italian. However, while it is true that neorealist films are part of the so-called dialect cinema, in order to make the narrative comprehensible and to reach as wide an audience as possible, the linguistic varieties present in these films are not faithful reproductions of the real communicative context of the time, but were adapted to the diegetic and stylistic needs of the director, creating a prefabricated language or, in the words of Raffaelli (1992: 107), an "imitative dialectality" 2 (translation mine). As Rossi (1999: 83; translation mine) observes: "The pretended dialectality of these films, in truth, is only a veneer that is limited to the most well-known and stereotypical features of dialectal pronunciation". 3 Therefore, in neorealist films, the prefabricated nature of orality in film language, the socalled "oralidad prefabricada" in Chaume's (2004: 168) words, is even more accentuated. In addition, as Ramos Pinto (2017: 6) notes, the "tension" that arises in the translation of nonstandard varieties lies not only in having to overcome linguistic problems, but also in the inherent pragmatic and semiotic difficulties resulting from the close relationship between the speaker, the medium, and the context in which the linguistic variety is used.
Numerous studies in the field of audiovisual translation have focused on the translation of linguistic varieties in audiovisual products since the 1990s (Agost Canós, 1998;Di Giovanni et al., 1994;Kovačič, 1995;Malinverno, 1999). More recently, the journal inTRAlinea devoted four special issues to this subject (in 2009, 2012, 2016, and 2020), titled "The Translation of Dialects in Multimedia". In the first three issues, the emphasis was on a restricted group of more or less codified minority languages that are not officially recognised, while the latest edition broadened the scope to deal with the translation of language variation from different perspectives. Some international research projects, such as TRAFILM, 4 focus on the challenge of translating multilingual audiovisual texts for the purposes of dubbing, subtitling and accessibility. Meanwhile, many studies have been conducted on audiovisual products to explore specific phenomena, including dialect-to-dialect translation (Ellender, 2015;Jäckel, 2001;Mével, 2007), the reception of translated humour (Antonini, 2005;Dore, 2017;Fuentes Luque, 2003), the representation of foreign languages on screen (O'Sullivan, 2011), the reception of translated cultural-specific references (Bucaria & Chiaro, 2007;Ranzato, 2016), and translation strategies to render linguistic variation (Corrius & Zabalbeascoa, 2011;Ramos Pinto, 2009, 2010 not only in films and television series but also in cartoons (Bruti, 2009;Minutella, 2020;Parini, 2019;Valleriani, 2020).
In the domain of subtitling, Hargan (2006), Petillo (2012), and Raffi (2017Raffi ( , 2021 are among the few scholars to have focused on the ways in which Italian varieties are rendered in the English subtitles of audiovisual fiction, using three different films as case studies, respectively: Roma, Città Aperta (1945) by Roberto Rossellini; Alessandro Piva's Mio Cognato (2003), andLe Notti di Cabiria (1957) by Federico Fellini-two of which are neorealist films. Given the contrast between standard and non-standard Italian (marked by regionalisms and stereotypical features of dialect), the authors sought to explore this dichotomy in terms of how it is reproduced in the subtitling process. It was found that sociolinguistic variation is generally reduced in English subtitles (Hargan, 2006;Petillo, 2012). However, thanks to the addition of spoken English and colloquial traits as well as non-standard grammatical forms in subtitles, dialect can be signalled in translation, and the retention of Italian words may serve to highlight the use of standard Italian, as opposed to dialect (Raffi, 2017(Raffi, , 2021.
As for Italian cinema, De Meo (2012) focused on director Crialese's Respiro (2002) and Nuovomondo (2007) to study the treatment of the vernacular oral medium in translation and to see how dialect is retained in the subtitles. According to her study, and in line with Raffi's (2017Raffi's ( , 2021, the transfer of features of spoken discourse in the translation as well as the use of nonstandard grammatical forms appear to be an attempt to maintain some of the features of the Sicilian dialect of these films. Later on, M.G. De Meo (2020) also concentrated on the detective television series Inspector Montalbano, which heavily relies on a mixture of Sicilian dialect and standard Italian, and found that the deletion of specific features of dialect (such as repetition and alteration of register) eliminated the sicilianità of the audiovisual product for the English-speaking audience. More recently, some of the contributions shared on the occasion of the conference Ragusa e Montalbano: Voci del Territorio in Traduzione Audiovisiva (Ragusa and Montalbano: Voices of the Territory in Audiovisual Translation; translation mine) and included in Sturiale et al. (2019) also dealt with Inspector Montalbano and the transposition of dialect through subtitling (among others, Bruti & Ranzato, 2019;Fiano & Martusciello, 2019;Sandrelli & Mecocci, 2019). According to these studies, indicators of dialect are in most cases standardised or neutralised in the corresponding subtitles, although colloquial forms are sometimes added to compensate for this loss.
Moving from Sicily to Tuscany, Bonsignori et al. (2019) focused on Paolo Virzì's culturespecific scenarios, which include regionalisms, slang, and colloquialisms. They identified the ways in which indices of regional and Italian identity are transposed in the English subtitled versions of three films, namely La Prima Cosa Bella (2010), Tutti I Santi Giorni (2012), and Il Capitale Umano (2014). According to their study, markers of orality in Virzì's works as well as the typical features of dialects and regiolects are mostly deleted, and the switch between Italian, regional Italian and local dialect is not reproduced. Nevertheless, the authors detected an attempt to preserve the lively nature of the original informal register by means of declarative questions and marked lexical choices.
Despite the above-mentioned valuable contributions, the research conducted so far is often limited to the analysis of the monomodal and textual dimensions of communication, overlooking the fact that the communication processes of audiovisual products are multimodal by nature. Therefore, building on the works of Iedema (2003), Jiménez Hurtado and Gallego (2013), Pastra (2008), and Ramos Pinto (2017), and adapting the model developed by Ramos Pinto and Mubaraki (2020), the present study focuses on the standard and nonstandard language varieties present in the original dialogues, i.e. the source text (ST), of Miracolo a Milano (Miracle in Milan), directed by Vittorio De Sica (1951) by Vittorio De Sica and in its English subtitles, i.e. the target text (TT). This film depicts the harsh post-war conditions of Italy, where orphanages were common. It tells the story of Totò, an orphan boy who settles into a homeless camp on the outskirts of Milan. When oil is found, a predatory landlord tries to displace its residents but Totò helps them thanks to a magic dove he has received as a gift. The extraordinary linguistic richness that characterises this film, which is certainly due to the work of screenwriter Cesare Zavattini, makes it an ideal case study to examine how linguistic variety is represented through audiovisual translation; in Zavattini's own words: "When writing a dialogue, I always think of it in dialect [. . .]. Using dialect, I feel it to be more essential, truer. Then I translate it into Italian, thus maintaining the dialect's syntax" (Sisto, 2014: 85). 5 The aim of this paper is to detect the overall translation strategy adopted in the target text as well as the intermodal relations established between the verbal and mise-en-scène modes in the ST and between the subtitles and the verbal and mise-en-scène modes in the TT. This two-level multimodal analysis (textual and diegetic), performed on an annotated corpus, 6 was designed to identify the different diegetic functions established by the intermodal relations present in the ST, making it possible to assess the impact of the adopted translation strategies in preserving, modifying, or even eliminating these intermodal relations in the TT. After this introductory section, Section 2 presents the study design including the construction of the corpus and the description of the analytical framework. In Section 3, the results of the analysis are presented and discussed. Finally, in Section 4, some conclusions are drawn.

Corpus
A bilingual parallel corpus composed of the transcription of the ST (the original dialogue in Italian) and the TT (the corresponding English subtitles) was built using the Blu-ray DVD edition of the film, distributed in 2015 by Arrow Academy, one of the UK's leading independent distributors of world cinema. The source-text component of the corpus was divided into units of analysis. In the present study, a unit of analysis is understood as a sentence intended as "a sequence of words initiated by a word in capital letters and concluded with a punctuation mark" (Ramos Pinto & Mubaraki, 2020: 9). Each unit of the source-text component was then aligned with the corresponding translation (i.e. the English subtitles). Finally, the units were organised into scenes, that is narrative episodes of variable length that take place in the same unit of space and time, delimited by an evident change in the communicative situation, 7 that is "distinct phases of the action occurring within a relatively unified space and time" (Bordwell & Thompson, 1979/2008. The corpus was then annotated using the UAM CorpusTool 2.8. 8 As for the textual level of analysis, two annotation schemes were created: one for the textual level of the ST and one for the textual level of the TT. For the present paper, due to space limitations, the analysis includes five scenes (see Table  1), which were selected for being particularly representative of different narrative moments of the film as well as of its linguistic richness (each scene will be thoroughly described in Section 3.2). The aim was to classify the units of analysis according to the language varieties used in the ST and TT, as further discussed in the following section.

Textual level
With regard to the TT, Ramos Pinto's (2017: 6) classification of the English language was adopted. A distinction was then made between standard varieties, to which a high degree of prestige and high sociocultural status are accorded, and non-standard varieties. Non-standard varieties were divided into the following categories: oral, regional, and substandard varieties. The latter was further organised into subcategories: regional, social, and social-specific. Each linguistic variety expresses and is associated with specific communicative meanings: the oral variety expresses a low degree of prestige, with a colloquial and informal use; the regional variety expresses a low degree of prestige and is localized to specific geographical areas; finally, the substandard varieties express a low degree of prestige and have a restricted use among specific social groups who have a low level of education.
As far as the ST is concerned, the model proposed by Ramos Pinto (2017) was adapted to the Italian language system, taking inspiration from the model of linguistic architecture (or linguistic repertoire) introduced by Berruto (2019: 24). An attempt was then made to elaborate a model of Italian based on a range of varieties, organised according to a prestige scale ranging from the most to the least prestigious. The degree of prestige of each variety was determined on the basis of the associated communicative meaning. Thus, the distinction was made between standard varieties, composed of linguistic features expressing a high degree of prestige, a high sociocultural status, a high level of education and representing the norm, and non-standard varieties formed by linguistic features deviating from the norm. The latter were further subdivided into the following categories: functional, which indicates a high degree of prestige, an informal register, a high socio-cultural status, and a high level of education; social specific, which is recognised as having a high degree of prestige, a formal register, a high socio-cultural status, a high level of education, and restricted use among specific social groups; and substandard. The latter was subdivided into oral, 9 which is associated with a low degree of prestige, a colloquial use, and limited to informal communicative situations; social regional, which conveys a low degree of prestige, a low socio-cultural status, a low level of education, and is particularly marked in diatopia; finally, social specific substandard variety, which expresses a low degree of prestige, a low socio-cultural status, a low level of education, and used by specific social groups.
By classifying 10 all the units of both the ST and TT in the corpus, it was possible to obtain frequency data showing how the distribution of these varieties changes between the ST and TT, revealing the overall translation strategy employed in order to eliminate or preserve in the TT the detected linguistic varieties of the ST.

Diegetic level
As for the diegetic level of analysis, the aim was to identify the intermodal relations established in the fictional world of the film between different modes 11 : spoken and mise-en-scène modes for the ST; speech, mise-en-scène, and subtitle modes for the TT. While the spoken mode includes lexical, phonetic, and morphosyntactic elements and the subtitle mode includes graphic, lexical, and morphosyntactic elements, the mise-en-scène mode includes setting, behaviour, costume, and makeup. Following Ramos Pinto and Mubaraki's (2020) analytical framework, the elements that make up spoken, mise-en-scène, and subtitle modes combine with each other and establish two core intermodal relations: confirmation and contradiction. Regarding the former, the meaning expressed by one or more signs confirms the meaning expressed by the others because they are semantically equivalent and/or are in line with what the viewer expects from the character. In this case, the presence of nonstandard varieties can perform several diegetic functions: to introduce a moment of authenticity and realism into the film; to indirectly represent the character, by indicating which social or regional group he/she belongs to, his/her level of education, and the specific historical period; to establish the interpersonal relationships of solidarity between characters belonging to the same social group; to offer the narrator's point of view if the character in question also plays the role of narrator. On the other hand, contradiction is established when the meaning expressed by one or more signs does not coincide with the meaning expressed by the others because they are semantically incompatible or opposite and/or are not in line with what is expected from the character. In this case, the presence of a non-standard variety serves to: differentiate the character from the others; introduce a moment of comedy, irony, or criticism into the narrative; establish interpersonal relations of power between the characters in cases where they do not belong to the same social group.
Therefore, an analysis of the diegetic dimension of the film makes it possible to see if the original intermodal relations of the ST are preserved, cancelled, or modified. Detecting the impact of translation strategies on subtitles is crucial because any changes made to the intermodal relations established in the source text may change the communicative meanings and diegetic functions of the subtitles, as further discussed in the following section. Table 2 shows the percentages of varieties identified in the ST, as well as the number of occurrences in brackets:

Textual level
Despite a slight general preference for the standard variety (51%) over the substandard variety (49%), there is a difference of only two percentage points. However, when we observe the distribution of varieties in each scene, we notice that the presence of the standard variety is higher than the substandard one only in two scenes out of five, in scenes 4 and 13. In these scenes, the standard variety is associated with the protagonist Totò and the rich industrialists Mobbi and Brambi. Despite the fact that Totò is an orphan and is therefore portrayed as a character with a low social status, the protagonist speaks mainly standard Italian because, compared to the other homeless people in the film, he has attended a few more years of school at the orphanage. The analysis of the diegetic level (see, Section 3.2) will indeed show that the standard variety associated with Totò has the function of accentuating the contrast with the other characters of the same social status. On the other hand, as far as the characters of Mobbi and Brambi are concerned, the standard variety reflects their social status and level of education, which, as we shall see, in Section 3.2, establishes a relation of contradiction with the environment and a relationship of power vis-à-vis the inhabitants of the slum.
By analysing the linguistic varieties associated with the other characters in the film, we come to the conclusion that, with the exception of the protagonist, the distribution of varieties faithfully reflects the social stratification of the characters. Standard Italian is used for characters with a high level of education and social status, such as Mobbi and Brambi in scene 4. In contrast, the social regional substandard variety is associated with characters with a very low social status and education, such as Alfredo in scene 5 and the slum dwellers in scenes 9, 12 and 13. In fact, the lowest typical Milanese traits were found in the speech of these characters, including the sonorisation of the intervocalic /s/ (e.g., Giuzeppe), the labialized pronunciation of the velar consonant (e.g., Con chi reumati chi se pò minga andà avanti), and the degemination of doubles (e.g., Ma deve sta atenta).
As regards morphosyntactic traits, we found: the adverbs of negation minga and mica in a postverbal position (e.g., L'è mica possibile andare avanti insci); the doubling of the pronoun particle lù (personal pronoun subject of courtesy) + el (personal pronoun in the third-person singular) (e.g., Podaria ospitarlo a mia casa, se lu el cred); the use of the article el as a personal pronoun, which becomes l' in front of a vowel (e.g., Ma a mì che me pias l'è la valis); and the use of dialectal morphological forms (bisognaria instead of bisognerebbe, podaria instead of potrei, demm instead of andiamo, etc.). On a lexical level, dialect terms appear, such as reumati for rheumatism, barba for delinquent, insci for so, and valis for suitcase. There are no instances of swear words.  It can be observed that in the TT the communicative meanings transmitted by the standard variety of the verbal modality are largely maintained (44%). Nevertheless, compared to the ST, the TT contains a higher percentage of the non-standard oral variety (56%) and this is mainly associated with the language used by the protagonist and the rich industrialists. From the analysis of the diegetic level (see, Section 3.2), it will be observed, in fact, that this choice entails the elimination of the diegetic function of differentiating the protagonist from the other characters who share the same social status. Meanwhile, with regard to the characters of Mobbi and Brambi, a new diegetic function will be seen to emerge, because the power relation established at the beginning of the scene becomes distorted.
It seems, then, that in the translation of the non-standard varieties of the ST, the same solution is chosen each time, namely the use of the non-standard oral variety. However, this means that the communicative meanings inherent to the verbal modality which are associated with the substandard regional variety are removed from the TT, since it does not convey information related to the geographical origins or social background of the characters, but simply signals the informality of the conversation.
It could be said that there was an attempt on the part of the translator to maintain the linguistic variation established in the ST, as shown in Table 4 However, when the distributions of standard and non-standard traits are compared between ST and TT, it can be seen that non-standard traits are more frequent in the TT: 56% against 51% in the ST. Consequently, it can be concluded that the overall translation strategy selected in the subtitle mode of the film is decentralisation, which entails an effort towards adequacy to the oral register of the ST and to the target cultural oral discourse (Ramos Pinto, 2017: 23). Furthermore, it should be noted that the selection of the non-standard oral variety in the TT signals to the viewer the informality of the communicative situation, but it is not able to elicit important information pertaining to the characters which is normally expressed by the use of this variety in the ST, i.e. their geographical origins, level of education, or social background.

Diegetic level
As described in Section 2.2.2, this paper recognises that the traits of the spoken (lexical, phonetic, and morphosyntactic) and mise-en-scène modes (setting, behaviour, clothes, and makeup) in the ST, and the traits of the spoken, mise-en-scène, and subtitle modes (lexical, graphic, and morphosyntactic) in the TT can be combined with the effect of establishing two types of intermodal relations: confirmation and contradiction. The results of the analysis of the intermodal relations of confirmation and contradiction are presented in Table 5 below.
Scene 4 is set in Piazza Baiamonti, near Porta Volta, outside the orphanage, where the protagonist finds himself homeless in a cold and unfamiliar city (i.e Milan). Totò wears the school uniform and carries a briefcase. He walks towards the centre of the square where, full of goodness and a will to live, he greets all the people he meets on the street. The standard variety associated with the protagonist establishes a relation of confirmation with his behaviour in order to perform the diegetic function of representing Totò as a character with a good soul. Incidentally, the film is based on Zavattini's novel Totò Il Buono which literally means "Totò, the good boy". With respect to the character's attire and the setting, Table 5 shows that the standard variety establishes instead a relation of contradiction, which fulfils the diegetic function of distinguishing Totò from the people he encounters by virtue of his high level of education, despite his low social status. However, in the dialogue with one of the passers-by, we notice that the protagonist switches from the standard variety to the same variety used by his interlocutor, i.e. the substandard social regional variety, characterised mainly by phonetic traits, including the presence of the vowels [ø] and [y]. This variation performs the diegetic function of establishing solidarity: it is associated with the passer-by and corroborates with the setting and behaviour depicted. Here, dialect undoubtedly serves to naturally portray a local character. However, the fur coat and elegant hat worn by the same character contradict the meaning expressed by the substandard variety. This particular relation of contradiction performs the diegetic function of differentiating the passer-by from Totò, who wears a modest suit.
In scene 5, a middle-aged man, poorly dressed, steals Totò's briefcase containing his only possessions in front of La Scala theatre. Despite an initial moment of hesitation, Totò plucks up the courage to chase the thief until he catches up with him in a dark suburban street. From the very first lines of their exchange, there is a clear alternation between Totò's Italian and Alfredo's Milanese. The standard variety establishes (along with the setting and the costume/makeup) a relation of contradiction to perform the diegetic function of increasing the contrast between the protagonist and the thief who have different levels of education. On the other hand, the substandard social regional variety and the setting establish a relation of confirmation in order to portray Alfredo as a local character with a very low social status and educational level. In the course of the conversation, Totò once again alternates between standard and substandard social regional traits in order to perform the diegetic function of establishing solidarity with the poor man. Faced with the man's reaction, Totò decides to give him the briefcase. By analysing the diegetic level of the ST, it can be observed that the protagonist's behaviour establishes a relation of confirmation with the standard variety in order to reiterate his charitable nature, a trait that recurs frequently throughout the film.
In scene 9, Totò and Alfredo walk around the slum to put up signs indicating the names of the streets. Here, they meet various inhabitants, including a very short old man intent on hammering a nail into the outer wall of his house, a man suffering from rheumatism, and a third man with a crooked eye and mouth. In the ST, the selection of the standard variety establishes a relation of contradiction with the environment and the clothing of the protagonist in order to perform the diegetic function of differentiating him from the poor by his level of education. In fact, in this scene we notice that Totò teaches them multiplication tables. Moreover, the alternation between standard and substandard social regional traits detected in Totò's speech establishes a relation of confirmation with his behaviour in order to perform the diegetic functions of indicating his desire to offer solidarity and representing him as a good character. In fact, when confronted with such characters, Totò adopts a whole series of behaviours to show empathy towards them: he bends his legs until he is the same height as the elderly man, and he also pretends to have rheumatism and imitates the face of the man with the crooked eye and mouth. Table 5 shows that the substandard social regional variety associated with the slum dwellers reinforces all the elements of the mise-en-scène mode in order to introduce a moment of realism and to indirectly represent the characters' social status.
In scene 12, Totò and Alfredo assign houses built in the slum to the homeless. A woman stands out among them, dressed in a way that is decidedly inappropriate. As mentioned before, she is wearing a hat and fur coat. In addition to the diegetic function of portraying Totò as an altruistic, sympathetic, and highly educated character, the substandard social regional variety associated with the woman establishes, together with her dress and haughty behaviour, a relationship of contradiction: this differentiates her from the other inhabitants of the slum and introduces a moment of comedy. Scene 13 shows a car driving along a dirt road leading from the town to the slum. Two men dressed alike get out of the car, both wearing a hat and fur coat. One is Mobbi, the wealthy industrialist, the other Brambi, the owner of the land on which the shacks stand. Following an ironic negotiation, Mobbi signs the contract and becomes the new landowner. The standard variety associated with the two tycoons establishes a relation of confirmation with their dress and behaviour in order to represent them as two characters with a high level of education and social status. The only exception is the setting, which instead establishes a relation of contradiction to reflect a relationship of power and to increase social diversity. As for the poor, the meaning associated with the substandard social regional variety confirms the meaning expressed by all the elements of the mise-en-scène mode in order to indirectly represent them as characters with low status and education.
By comparing the intermodal relations established in the ST and TT, it is possible to assess the impact of translation strategies in preserving, modifying, or even eliminating the intermodal relations originally established in the ST.
Turning our attention to scene 4, it can be observed that the non-standard oral variety associated with the passer-by in the subtitle mode maintains the relation of confirmation with the behaviour and setting established in the ST (see , Table 5). However, Table 6 shows that the meaning associated with the non-standard oral variety cancels out the contradiction with the way the speaker is dressed, thus eliminating the diegetic function of differentiating the passer-by from Totò. Furthermore, the use of the non-standard oral variety in the subtitles eliminates the diegetic function of representing the passer-by as a local character: it is possible to infer the informality of the conversation from this variety, but not the character's origins or specific social group to which he belongs. With regard to the protagonist, it is evident that the diegetic function of representing him as a highly educated figure is eliminated in the TT as Totò's speech contains more traits belonging to the non-standard oral variety than the standard one. Consequently, on the basis of the linguistic variety selected in the subtitle, the target culture viewer can only infer the informal nature of the conversation, whereas the speaker's geographical origins will have to be guessed from the title of the film and his social status from the elements of the mise-en-scène mode.
In the fifth scene, in the TT, the diegetic function of differentiating Totò from Alfredo is maintained (see , Table 5) thanks to the selection of the standard variety that contradicts the costume/ makeup and setting. Similarly, the linguistic variation in Totò's speech remains unchanged because elements of the translation belong to a less prestigious variety, namely the non-standard oral category. Moreover, the behaviour and the standard variety of the subtitle and spoken modes maintain the relations of confirmation established in the ST (see, Table 5). As regards the character of Alfredo in the TT, the selection of the non-standard oral variety is in line with the spoken and mise-en-scène modes. In spite of this intermodal relation of confirmation, it should be noted that the non-standard oral variety in the subtitles does not convey either Alfredo's geographical origin or his association with a specific social group. These meanings will only be communicated by the mise-en-scène mode and the film title.
Turning to scene 9, Table 6 shows that the original corroboration between the standard variety and behaviour (see , Table 5) is maintained in the TT by the choice of the standard variety in the subtitle mode. Meanwhile, the intermodal contradiction established in the ST between the standard variety, costume/makeup, and setting serves to accentuate the contrast between the protagonist and the old man he meets in the slum. In the subtitle mode, the standard variety is attributed to Totò and a nonstandard variety is accorded to the old man, and these choices do not incur any modification to the diegetic function established in the ST (see , Table 5). Another diegetic function that is maintained in the TT is the relationship of solidarity between Totò and the inhabitants of the slum. In fact, the way in which Totò's speech switches between the standard and substandard regional varieties is reproduced in the TT by replacing the standard variety with the non-standard oral variety. Furthermore, from Table 6 it can be seen that in the TT the non-standard oral variety together with the substandard regional variety of the spoken mode are in confirmation with all the elements of the mise-en-scène mode to preserve the diegetic function established in the ST of introducing a moment of realism and indirectly representing the poor status of the inhabitants of the slum. As previously mentioned, despite the fact that the non-standard oral variety does not allow the target viewer to understand the geographical origins and social status of the poor, it is possible to intuit their profile thanks to the meanings expressed by the mise-en-scène mode.
In scene 12, as shown in Table 6, in the subtitle mode the selection of the standard variety associated with the main character maintains the diegetic functions established in the ST (see ,  Table 5). Similarly, the linguistic variation found in Totò's speech in order to establish solidarity with the inhabitants of the slum is maintained in the TT by the presence of traits belonging to a less prestigious variety. Moreover, Table 6 shows that in the TT the selection of the non-standard oral variety maintains the relations of contradiction established in the ST (see , Table 5) between the substandard variety, costume/makeup, and haughty behaviour assumed by the woman, all of which perform the diegetic function of differentiating her from the other poor people and introducing a moment of comedy.
In scene 13, as shown in Table 6, even if all the diegetic functions established in the ST (see , Table 5) are maintained in the TT by the choice of the standard variety, the TT contains a higher percentage of non-standard oral traits associated with the figures of the rich industrialists, introducing a new diegetic function: the power relation established at the beginning of the scene is cancelled out, and there is a shift in the TT to the same variety associated with the poor people. Consequently, the reader of the subtitles could be (mis)led to believe that both Mobbi and Brambi want to establish a relationship of solidarity with the slum dwellers, a function which is totally absent in the ST.

Final remarks
The language of Miracolo a Milano, founded on the compromise between mimicry, synthesis, social criticism, and comprehensibility, is characterised in the film through the selection of mainly phonetic traits. We found that the most obvious features of Milanese are the sonorisation of intervocalic /s/, the labialised pronunciation of the velar, and the degemination of doubles. This variety raises several questions from a translation point of view as it is closely linked to the speakers, the communicative medium, and the context in which it is used, and therefore evokes specific communicative meanings. In fact, the challenge posed by the translation of non-standard varieties in films does not so much lie in the transposition of oral discourse into written discourse, nor in the translation of the non-standard linguistic varieties per se, but rather in finding linguistic varieties within the target culture that have the same associated communicative meanings and are thus able to maintain the same diegetic functions. Table 2 shows that the non-standard oral variety is the only non-standard variety present in the TT. This result confirms the findings of previous studies: there is a tendency to compensate for the loss of information associated with regional non-standard varieties by using oral traits to mark speech as non-standard. The non-standard oral variety was recreated in the TT by the use of mainly graphic features, such as contractions. In addition to graphic elements, typical features of oral speech were found such as the absence of subject-verb inversion in questions, ellipses in questions and answers, and discursive cues including interjections. It is assumed that the translators selected mainly graphic features to ensure that the subtitles were easily readable and accessible to a wider audience. However, it can be concluded that the use of the non-standard oral variety in the TT, as the only solution to translate the non-standard varieties of the ST, eliminates the communicative meanings associated with the regional substandard variety as it fails to convey information about either the geographical area of origin or the social background of the characters, signalling only the informal register of the conversation.
The quantitative analysis carried out at the textual level allowed us to compare the frequency percentages of the standard and non-standard linguistic varieties present in the ST and TT in order to identify the overall translation strategy employed in the subtitles (see , Table 4). Generally speaking, it is apparent that there was an attempt by the translators to maintain the linguistic variation of the ST by selecting both standard and non-standard varieties in the TT. In particular, we found that the translation strategy selected is one that favours decentralisation, since nonstandard traits are more frequent in the TT: 56% as opposed to 51% in the ST.
Tables 5 and 6 reveal how different relations of confirmation and contradiction are established in the ST and TT to perform specific diegetic functions, such as introducing realism into the film, indirectly representing the characters, and establishing interpersonal relations of solidarity and power. Once the diegetic functions of the ST and TT were identified, it was possible to evaluate the impact of the translation strategies in preserving, modifying, or eliminating the intermodal relations established in the ST. In fact, any modification of these was shown to introduce diegetic functions to the TT not foreseen in the ST.
In neorealist films in general, and in certain films that can be related to the movement, nonstandard varieties are used not only to faithfully represent the pluralistic world of post-war Italy but also to create a binary opposition, positive versus negative: on the one hand, the positive code represented by lower-class Italians and associated with non-standard Italian; on the other, the negative code represented by middle-class characters speaking in standard Italian. As for Miracolo a Milano, in the TT the standard variety is associated with the wealthy characters on the one hand, and the non-standard oral variety is associated with the social outcasts on the other. These choices maintain the intermodal relations and diegetic functions established in the ST. In some cases, however, we found that the selection of the non-standard oral variety in the subtitles eliminates or modifies the diegetic functions established in the TT (see , Table 6). For example, in scene 5 in which the protagonist meets the homeless man Alfredo for the first time, we noticed that the selection of the oral non-standard variety in the subtitles fails to convey Alfredo's geographical origins or adhesion to a specific social group, eliminating the diegetic function of indirectly representing the character. These meanings will only be conveyed by the title of the film and the resources of the mise-en-scène mode. Similarly, in the first scene, set outside the orphanage, it was established that in the subtitles the meaning associated with the non-standard oral variety incurred the elimination of the relation of contradiction with the way in which the passer-by is dressed, thus removing the diegetic function of differentiating him from Totò. In the same scene, it was also found that the diegetic function of representing the protagonist as a highly educated figure was deleted as Totò's speech takes on more traits belonging to the oral non-standard variety than the standard variety. In scene 13, we also observed that the TT contains a higher percentage of non-standard oral traits associated with the figures of the rich industrialists compared to the ST, introducing a new diegetic function, i.e. cancelling the power relationship established at the beginning of the scene and misleadingly suggesting a desire to show solidarity with the poor people.
In conclusion, given the complex semiotic system of films, it is essential to focus not only on the verbal components, but also on the relationships established between the different modes at play in order to preserve linguistic variation which, in neorealist films, also creates a binary opposition between positive versus negative (i.e. standard and non-standard Italian). Having said this, it is not always possible to find a linguistic variety in the target culture that expresses the same associated communicative meanings expressed in the ST, and in this case the most appropriate translation strategies should be selected to maintain the communicative meanings and diegetic functions established in the ST as far as possible.

Funding
The author received no direct funding for this research.