In recent years, the interest in song translation has grown and several studies have been conducted in this field (Gorlée 2005; Kaindl 2005; Susam-Sarajeva 2008; Minors 2013; Low 2017). Different genres are analysed (such as opera, films, folk music, cover songs), but research on stage musicals is still sporadic (Franzon 2005; Mateo 2008). Some of these scholars developed models of translation for songs in performance, but they tend to focus solely on how to treat the lyrics (Low 2003; Franzon 2005), and not on the interaction between the different semiotic modes at the core of this genre. When the focus is on this multimodal aspect, it is either looked at in relation to opera (see review of Kaindl’s 1995 work by Clüver 2008) or presented as an opportunity for further research, for instance a “stage effectiveness” criterion to be potentially added to the Pentathlon Approach (Low 2017: 110).

According to Borroff (1984: 101), a stage musical is “a set of musical numbers conceived as one score with a balanced aesthetic whole.” In a stage musical, songs are inserted in a coherent narration that follows the scheme “beginning-development-conclusion,” and they allow for the advancement of the plot. Songs are “storytelling element[s]” (Kenrick 2008: 95) and represent the focal points of the performance. In stage musicals dialogues are important, but they do not necessarily convey the most relevant information, which is left in large quantity to the songs. In Woolford’s (2012: 245) words, “[s]ongs [are] capable of forwarding plot, developing character, commentating on the action or otherwise functioning in the story telling.”

The term stage musical refers to a live on-stage performance, of which several types can be distinguished; musical theatre is a broader expression used to define this genre (Woolford 2012: 20).

The available translation models that could be applied to songs from stage musicals are the Pentathlon Approach devised by Low (2003, 2005) and the functional approach based on communicative clues devised by Franzon (2005).

Low’s Pentathlon Approach (2005: 185), also known as the “pentathlon principle” (2003: 92, 101-102; 2005: 191-192) was not developed specifically for songs from stage musicals, but it provides useful suggestions on the translation of songs in general. At the heart of this approach is the idea that compromises are necessary, and that it is the translator’s task to decide which aspects to prioritise. The Pentathlon Approach is based on five categories that the translator should consider when approaching songs: singability, sense, naturalness, rhyme and rhythm. Singability is the quality that makes a song performable in translation. Sense refers to the general message of the source text that is to be reproduced in the target text, while naturalness is linked to the language used, which has to sound natural to the target audience. Sense and naturalness are relevant for the study of songs; however, they seem too vaguely sketched to be of use in a systematic analysis of these texts. What seems to be missing is a clear definition of their nature and the proper way to assess them in a research context. Rhyme and rhythm are related to the musical aspect of songs, to their style, and to the constraints that the music poses to both the songwriter and the song translator. Depending on the type of song, these issues will have more or less relevance for changes in the semantic meaning.

Several guidelines can be found on how to approach a song from a rhythmical point of view, but fewer exist on similar semantic analyses. While both Low and Franzon hint at the importance of the content of a song, it would appear that the reproduction of the semantic meaning often implies compromises, given the constraints that the hybrid nature of songs entails.

In this paper, the “essence of the source text” (Low 2005: 184) is indicated with the term message, used to refer to the content of the song. Message indicates what is expressed, but also any possible extra-textual information that is expected to be inferred from that content. Borrowing from musical terminology, the message is what Woolford (2012: 258, 260) defines as the “hook” of a song, which describes “a neat idea that captures the essence of the moment.”

Low devises a scoring system for rhymes, but does not do the same for the criteria of singability, sense, naturalness and rhythm listed in his model (Low 2008: 12); therefore, there does not seem to be a clear and systematic way to assess the transfer of meaning from the source to the target text. It is evident that, when translating a song, strict fidelity cannot be obtained and that some liberties have to be taken, if only for reasons of musicality. In relation to opera, for example, Desblache (2013: 16) states that “[o]pera translators’ priority is not semantic fidelity.” Moreover, in Franzon’s (2005: 292) words, “the verse form, partly determined by the structure of the music, is generally preserved, but the semantic content of the source lyric is handled very freely [in a process of] functional re-interpretation.” What emerges from the Pentathlon Approach is that a certain degree of manipulation is permitted (Low 2003: 94), and that the task of translating a song needs constant reworking.

While lacking from Low’s Pentathlon Approach (2003; 2005), function is a fundamental part of Franzon’s model (2005): a song translated into a target language needs to replicate the function of the original source text, rather than just its textual properties. The idea that songs have a specific function has been addressed, among others, by Zumthor (1983/1990: 143, translated by Murphy-Judy) and Laliberté (2005).

While not specifically clarifying what he means by “function” of a song, Franzon relates it to the concept of “functional units/communicative clues” (Franzon 2005: 263, 265-266), borrowed from Nord (1997) and Gutt (2000). These communicative clues are inserted into a functional model and found “in the properties of playwriting: character, conflict (interpersonal address), plot, milieu (context), and language (code)” (Franzon 2005: 274). This first definition does not seem to specify what communicative clues are, where to find them or what exactly this expert means with the expression “properties of playwriting.” He then adds that,

[f]unctional units may be found in pronouns, verbal modes, utterances implying emotion or attitude (interpersonal address), deictic or spatiotemporal references, mention of props (context), and implied staged activity, including gestures or verbal behaviour.

Franzon 2005: 274

Despite this addition, it is not entirely clear what type of relationship exists between the categories that the researcher identifies and the communicative clues that these categories seem to include.

In relation to the message of the song, Franzon (2005: 282) affirms that “translators often make large or small adaptations of culture-specific items.” This concept of acculturation is particularly relevant when it comes to theatre texts because, in Aaltonen’s (1996: 204) words, “[they] seem to tolerate substantial inconsistency in detail, as the audience does not have the time or need to reflect on details.” The same could be said of a stage musical, where there is no time for the audience to re-listen to a song, or to look at texts during the performance.

This overview of available models of translation aimed at songs demonstrated how the audio and visual modes are not as thoroughly analysed as the verbal one. The discussion started by Kaindl (2005: 235) on the “plurisemiotics of pop songs” and opera needs to be expanded to musical theatre as well. This means that there is a need for a specific model of analysis that would cover all three semiotic modes at play in songs from stage musicals (verbal, audio, visual). The identification of this gap has led to the development of the multimodal model of analysis for songs from stage musicals described in this paper.

The model was applied to a corpus that included songs from various stage musicals. For the purpose of this article, two stage musicals were selected to provide examples: Cats[1] and Jesus Christ Superstar (JCS).[2]

A stage musical tells a story in which characters interact on stage as if they were in the real world. If in everyday conversations people tend to follow specific patterns, then there are likely to be conversational rules in place in theatrical communications as well. However, there are differences between the ways in which communication proceeds on stage as opposed to real life.

First, the communication is not real, but carefully crafted, because the interaction was decided by the authors of the script; the spontaneity typical of real-life conversation cannot be replicated nor expected. Second, the communication takes place mostly via songs, rather than through dialogues or monologues. It is therefore necessary to interpret songs as acts of communication carrying significance and a communicative value.

Songs are considered acceptable forms of communication used by the characters on stage to convey information and emotions between themselves and with the audience. Music, not normally present in everyday conversations, does not stop these exchanges from happening; on the contrary, all the participants accept it. The context in which the communication takes place is composed of characters, plot, place and time of the story, and by audience, and place and time of the staging.

Traditional rules of conversation seem to be transposed to songs, which act as monologues and dialogues, and despite their “prefabricated orality” (Baños-Piñero and Chaume 2009) they are analysed as regular conversations. Even though it was previously crafted by the songwriter, the interaction based on songs is expected to be accepted by the audience not just as spontaneous, but also as the norm.

From a pragmatic point of view, songs have a “spatio-temporal characterization” (Van Dijk 1977: 191) that refers to the time and place of narration: the way in which these elements are referred to in the text is an indication of how the context impacts the text itself. There may be situations in which some references are left unexplained or given as granted, drawing from knowledge that the audience is supposed to possess. In addition to that, Lewis’ (1972) co-ordinate of “previous discourse” allows “the hearer to interpret what is said in the light of what has already been said” (Brown and Yule 1983: 41) and is related to the idea of a shared knowledge created throughout the performance. Relevant concepts and references might not be repeated or explained every time, but just hinted at; in this case the hearer, and the audience with them, can grasp the meaning based on what was previously said. In this pragmatic theoretical framework, shared knowledge can also be related to the “intended meaning” defined by Cruse (2004: 22), that is those elements that do not need explanation in order to be understood by the hearers and the audience. Shared knowledge is established both between the characters on stage that act in the fictional story, and between the characters on stage and the audience, which gets involved in the information provided while attending the performance. On the concept of existing knowledge, Oittinen (2008: 86) states that “readers need to fill in information gaps found in texts on the basis of their background knowledge as well as their experience.” This is relatable to stage musicals, where the audience is often expected to fill in “gaps” throughout the performance. This means that translators approaching songs from stage musicals should consider the degree of knowledge that the target audience is expected to possess and consider the eventuality of having to make concepts more or less explicit.

The purpose of the multimodal model of analysis for songs from stage musicals presented in this paper is to approach a less-researched genre, offering a new perspective on the way songs in stage musicals could be analysed and translated. The methodology used for the analysis was created by assembling already existing theories that refer to verbal, audio or visual modes separately; multimodal approaches have started to develop only recently, and a univocal framework is still to be produced.

This model looks at the “semantic interaction among semiotic systems” (Pastra 2008: 300) and aims to provide more information on the multiple modes at play in songs of stage musicals, which could eventually inform their translation. It was developed with a particular focus on the identification of relevant themes in the lyrics, which are then replicated (or not) in the audio and visual modes. These themes represent the information that each song wants to transmit to the audience, and they are expressed via verbal, audio and visual semiotic resources.

Songs could be defined as “ensembles, representations or communications that consist of more than one mode, brought together not randomly but with a view to collective and interrelated meaning” (MODE 2012).[3] This definition also fits the concept of themes used in this paper; according to how many different semiotic resources are employed, a theme will be considered as contributing more or less strongly to the overall meaning of a song.

As stated in MODE (2012), “the weighting of modes, in terms of which [they are] predominant or backgrounded, can indicate relative ‘status’ and their very co-presence is suggestive of multimodal interrelationships.” The concept of status was transformed in this research into multimodal richness, based on the different types of themes that can be classified. A theme that is expressed by a high amount of different semiotic modes contributes more to the message of the song, and it is more likely to be maintained in translation.

Table 1 presents an overview of the model and of the semiotic resources that will be discussed next.

Three types of relations were identified between the verbal, audio and visual modes: addition, enhancement and modification (see Table 5). These categories focus on the way the audio and the visual modes contribute (or not) to the dissemination of the meaning of the verbal mode, and on how they complement the themes that were identified in the text.

These categories were created drawing from Halliday’s (1994) logico-semantic theories, in addition to the work of Barthes (1964/1967), Martinec and Salway (2005), Pastra (2008) and others, and were modified according to the purpose of this study. Addition, enhancement and modification bring together existing theories by linking verbal, audio and visual modes, and include both the relationship between verbal and audio modes and the relationship between verbal and visual modes.

The category addition signals the introduction of a new theme by one of the semiotic modes at play in a song (verbal, audio and visual). As previously explained, the themes are identified via specific categories designed to extract information and labelled using key words covering a vast area of meaning. When a theme, usually expressed in the verbal mode, is not replicated in any other mode, a blank box is used to signal the omission.

The category enhancement draws from the categories of “extension” and “enhancement” defined by Martinec and Salway (2005). It identifies situations in which themes that were first located in the verbal mode are replicated in the audio and/or visual modes, with extra details added to them. This also shows how the semiotic modes collaborate with each other in a song, whose message is therefore not expressed solely by the lyrics. In Table 6, the theme of agility is enhanced via the audio and visual modes, which reinforce this interpretation and add information to it. This example also shows how two of the themes expressed verbally are enhanced via audio resources, while three are enhanced visually. Three new themes expressed by the audio mode are also enhanced visually.

The category modification takes an existing theme and provides an additional interpretation for it: the theme is slightly modified, but not enough to become a new one. For this reason, the theme does not stand on its own as new, but is linked to another one; a different label is provided in the table, but the original one is also maintained.

Table 6 provides the themes identified in “Jellicle Songs for Jellicle Cats” (Cats) and defines the relationships between the modes. As can be seen in the first column of the table, four new themes are added via verbal mode, three via audio mode and one via visual mode.

The column labelled Modality refers to the number of modes that are used to express each theme (multimodal richness): “multimodal” (three modes), “bimodal” (two modes) and “monomodal” (one mode). This classification is based on the idea that more concepts are conveyed when a higher number of different semiotic resources are used; a hierarchy of themes is established in each song, helping to focus on the more important concepts.

The aim of this paper was to fill an existing gap in the field of translation studies and to develop a model for the analysis and potential translation of songs from stage musicals, whose message is expressed by verbal, audio and visual semiotic resources. This paper showed not only that verbal semiotic resources express a relevant portion of the message of a song, but also that audio and visual semiotic resources can enhance, modify and introduce new themes, which should then be addressed when translating the show into a target language. Each step of the analysis was informed by the previous one, showing how verbal, audio and visual resources interact to create meaning.

This is the first time that the concept of themes has been used to explore the semiotic content of songs from stage musicals. “One/two word” definitions were employed to highlight relevant information expressed by different modes. Looking for themes structured the analysis of the songs and immediately pinpointed relationships of enhancement and modification. Focusing on the identification of themes is not only useful to understand the message of songs, but it can also provide a systematic path for the analysis of each semiotic mode.

Further attention will need to be devoted to the development of a terminology to be applied to the description of multimodality in performance. So far, terms have been borrowed from different domains, and a coherent and inclusive glossary is still being created. Such a terminology should not only be used to describe texts openly characterised by more than one semiotic mode (such as songs of stage musicals), but also to explore the potential of texts that could become such (for instance, poetry readings). For example, theatre plays may benefit from such an outlook, especially when audio and visual elements are involved.

Looking at songs from a multimodal point of view facilitates a more comprehensive understanding of their message and highlights which areas of the song are more open to interpretation than others. This is why knowing how to deal systematically with song translation could improve this process, for example by developing new strategies to deal with specific translation issues (like issues of cultural background) and possibly boosting the adaptation of new shows into different target languages. It could be a starting point for comparisons of different stagings of the same stage musical, of stage musicals created in different periods, and of stage musicals in different languages. This would also contribute to the growing field of research into re-adaptation and retranslations.

To conclude, a structured approach to the analysis of songs from stage musicals has offered informative insights into their complexity, showing how verbal, audio and visual modes interact to create meaning. Songs have proved to be more than just their lyrics, and this multimodal view could not only help professional song translators, but also be used to support the teaching of song translation itself.