Normal View Dyslexic View
Research Article
Engaging with art song in lesser represented languages: expanding or challenging the linguistic regime?
Eva Moreda Rodríguez*†
eva.moredarodriguez@glasgow.ac.uk
Abstract
While the transnational and translational ramifications of art song have started to attract attention in recent years, there is little critical understanding of the unique modalities in which multilingualism expresses itself in the genre. In this context, this article has a two-fold aim. Firstly, drawing upon scholarship both on art song and Multilingualism Studies, I map out what I call the linguistic regime of art song, defined by the multilingual competences expected of individuals active in the art song ecosystem (singers, composers, translators, coaches, audiences) and the hegemony of certain languages over others. Secondly, on the basis of an empirical experiment involving practice-led research in singing, I consider how engagement with songs in lesser represented languages might be encouraged, and how such engagement might help expand and/or challenge some aspects of the linguistic regime of art song.
Keywords
multilingualismart songpractice-led researchsingingIntroduction
In the art song sphere, multilingualism and monolingualism coexist in curious ways. If we pick up a few art song recital programmes at random from the thousands that are performed in public every year around the world, it is likely that a significant number of them will be multilingual, comprising songs in different languages.1 The singers performing these will likely have had multilingualism embedded in their training: conservatoires and other educational institutions across the Western world feature repertoire classes in different languages as a compulsory graduation requirement,2 and auditions might specifically require that candidates bring pieces in a number of different languages.3 At the same time, however, other aspects of the art song scene reflect some of the ideologies of monolingualism that accompanied the rise of the nation-state: multilingualism within the same song or even song cycle is rare, which makes it easier to classify them into separate linguistic-national traditions for the purposes of programming and teaching (‘French art song’, ‘German art song’, ‘Russian art song’, etc).4 Similarly, language training for singers tends to rely on a highly uniformised pronunciation standard, which often leaves little room for dialectal or otherwise non-normative pronunciation—a feature also typical of nationalism-led language standardisation throughout the 19th century and beyond. Besides, multilingualism in art song tends to be limited to a small number of hegemonic Western languages: French, English, German, perhaps with the addition of Russian, Czech, Italian and Spanish—and often the singer’s native language, if different from these.5
While numerous initiatives (anthologies, diction guides, online directories) have appeared in the last two decades, which attempt to provide singers with the resources and tools to engage with an art song repertoire in lesser represented languages than the seven named above, in this article I argue that such interventions—past, present and future—must be contextualised within what I call the linguistic regime of art song. In the first half of the article, I endeavour to map out this regime by examining existing bibliography both in art song and in the field of Multilingualism Studies. In the second half of the article, I present a recent practice-led experiment aimed at identifying some of the strategies that singers put into practice when learning an art song in an unfamiliar (to them) language. In what is an underexplored topic, I use the results from this experiment as a pilot to achieve two aims. Firstly, I aim to illuminate further aspects of this linguistic regime that might not be obvious from an analysis of institutions, discourses and repertoires, and that necessitate engagement with practice as it happens. Secondly, I discuss the strategies that came to light during this experiment as a starting point to pose the question of how multilingualism can be expanded in the art song ecosystem—whether in additive ways (that is, facilitating access to resources in new languages, but without fundamentally changing the genre’s practices), through challenging of some aspects of this linguistic regime or through a mixture of both.
From translation to multilingualism
The issue of language (of multiple languages) in art song has timidly started to emerge as studies in the genre have increasingly centred its transnational, mobile, performance-mediated nature (e.g., Loges & Tunbridge 2020). Here, attention has been focused mostly on translation. This, on the one hand, is consonant with earlier studies on opera, popular music and art song itself (Low 2006; Susam-Sarajeva 2008; Mateo 2013); on the other hand, it draws upon new insights in the intersection of Translation Studies and performance (Marinetti & De Francisci 2022; De Francisci & Marinetti 2025). In practical terms, this means that recent work has moved beyond a rigid notion of text, under which translation is understood primarily as the passage of song lyrics from one language into another (with rhyme and syllabic adjustments as needed to enable singability), and it now considers a broader range of practices having to do with adaptation of various kinds across multiple languages, territories and sometimes media (e.g., as in the chapters in Bullock & Tunbridge 2021). While still keeping a considerable focus on the notion of translation, some work has also displayed innovative approaches, such as asking how audiences engage with text in multiple languages during the concert experience (both through translation and otherwise) (Campbell 2023), and how composers tackle the composition of art song in different languages (sometimes through translation, sometimes not) (Gamrat 2023).
I would like to argue, however, that translation, no matter how broadly understood, is only a subset of the multilingual practices that conform the linguistic regime of art song: as in the example above, a composer might use a translation or engage in translation when setting text in multiple languages (for example, to create different versions of the same song, or to understand text in a language she does not speak well), but this does not always need to be the case: for a multilingual composer (such as Robert Gerhard, a Catalan composer of Swiss parentage, who wrote songs in German, Catalan and Spanish), the setting of texts in multiple languages does not need to involve obligatory translation. Similarly, a multilingual art song recital might not involve translation per se (the various songs will not normally be translations of one another, but rather completely different sets with different texts), but, throughout it, the singer and her audiences would have been exposed to multilingualism in a way that seems relevant to understand how the experience of art song unfolds. I therefore argue that expanding the focus of enquiry from translation to multilingualism can give us a more holistic understanding of how language regimes unfold in art song and indeed constitute a fundamental part of how the genre is configured, requiring the various participants (writers, composers, translators, singers, audiences) to engage with individual languages and with the notion of multilingualism itself in different ways.
Indeed, whereas I do not regard Musicology or specifically art song scholarship to be entirely ‘language-indifferent sectors’ (Footitt 2022): 218), I agree with Footitt in her assessment that many disciplines with an allegedly international focus (as increasingly is the case for Musicology, most notably via the rise of ‘global musicology’ in the last few years) have not sustainedly engaged with multilingualism, and would benefit from doing so—particularly at a time in which the field of Multilingualism Studies itself is expanding and becoming theoretically richer. There are, in particular, three key aspects of bibliography in Multilingualism Studies which I regard as particularly useful to understand how multilingualism determines linguistic regimes in art song. The first, and perhaps most obvious, is the development of the notion of ‘creative multilingualism’ to study the interactions between multilingualism and the creative arts with a focus on how the presence of or engagement with multiple languages impacts on individual and group creativity (Kohl & Ouyang 2020: 1–2). While the notion of a ‘regime’ of art song might imply that individual creativity is somehow curtailed—as participants need to use languages in specific ways—the practice-led component of this project revealed practices, strategies and behaviours from singers which can regarded to be as creative ways of adapting to this regime, which would likely never have emerged if participants had not been required to engage with multilingualism.
A second key takeaway from literature on multilingualism is the distinction between multilingualism at individual level (a multilingual individual is someone who knows more than one language) and at societal level (a multilingual society is one in which two or more languages are routinely present) (Ayres-Bennett & Fischer 2022: 3). Switzerland, for example, is a multilingual state with four official languages (German, French, Italian and Romansh), but individual citizens will rarely be multilingual in all four languages: most will speak the language of their region, and perhaps another of the country’s official language. Moreover, they might also speak English and/or a heritage language. The distinction between individual and society is useful in the case of art song: even if particular individuals within the art song ecosystem are monolingual (say, a composer who only sets text in her native language), the ecosystem they are part of is not: the songs by this composer, for example, might be part of a multilingual recital programme, and hence sung by a singer with some multilingual skills and heard by audiences alongside songs in other languages. Therefore, multilingualism, while not directly present in the song or in its composition process, can play a part in the performance and reception history of that song.
Thirdly, while earlier studies of multilingualism (or rather, bilingualism) often established that a certain measurable degree of competency in two or more standardised languages was necessary in order to consider an individual or a society to be multilingual, recent work has been considerably more critical of such notions of correctness, boundedness and competency. The notion of languages as ‘enumerable and bounded entities’ has itself come under scrutiny (McKinney et al. 2023: xxii), and, in some contexts, what might have been regarded earlier as dialectal variation is now being studied from the perspective of multilingualism. Crucially, there is also greater understanding of the fact that multilingual individuals might not know all their languages to the standard of proficiency that is typically assumed of an educated native speaker, and they might even have uneven levels of ability across the four skills (reading, writing, speaking and listening) that are typically regarded to conform language proficiency. Instead, scholars prefer to talk of ‘repertoire(s) of styles and linguistic resources’ (McKinney et al. 2023: xxiii) or ‘unit(s) of communication that transcend individual languages’ (Lo Bianco & Aronin 2000: 6) to refer to the individual, adaptative and in some cases highly creative ways in which multilingual speakers might make use of their languages. These strategies also include practices now extensively studied in modern literature on multilingualism but previously neglected as non-normative, such as translanguaging and code-switching (Lewis et al. 2012). Such asymmetries in individual language competence are also commonly found in the art song ecosystem, and it is fair to say that they happen in ways that are virtually not found anywhere else ‘in real life’, which makes such practices worthy of study through the prism of multilingualism. For example, a singer might have perfected his French pronunciation to levels nowadays not typically expected even at a C2 exam of ‘Français Langue Étrangère’, but he might not be able to sustain a short conversation at a hotel check-in in Lyon. Similarly, a composer wanting to set a German poem to music might develop intricate a level of knowledge of the metric and prosody of the language (or its 19th-century variant), which most natives will be nowhere near possessing, but she might not be able to read and understand the menu at a restaurant in Munich.
Importantly, levels of individual versus collective multilingualism, and levels of competence within individual languages will not depend purely on individual preferences or experiences outside the music realm: instead, they will be strongly shaped by a range of expectations regarding both language competence and attitudes towards language, which I collectively call the ‘linguistic regime’ of art song. These expectations are conveyed and upheld through cultural and educational institutions, and might be implicit or explicit. An explicit expectation, for example, could be a conservatoire or university department requiring candidates to sing pieces in given languages at their audition (so that a singer not meeting this requirement would be immediately disqualified and prevented from continuing their training). A more implicit expectation, for example, could be the songs that a composer wrote in a language non-native to them not becoming part of the repertoire because of perceived shortcomings or ‘unnaturalness’ in the text-setting. Table 1 attempts to summarise these expected competencies and attitudes; while it does not purport to cover every situation, it draws upon recently published work in the area of translation in art song, and on what we might term tacit knowledge embedded in the classical music industry as well as education institutions (conservatoires, universities).
Table 1
The linguistic regime of art song.
A table that describes the levels of linguistic competence expected of different participants in the art song ecosystem.
| Participants in art song ecosystem | Expected language competencies and attitudes |
|---|---|
| Composers | Understanding of metric structures, accentuation and prosody of the language; understanding of meaning (Binder 2021; Gamrat 2023). |
| Translators | Advanced passive competence in the source language; knowledge of metric structures, accentuation and prosody in the target language, and how they adapt to music. |
| Singers | Native or near-native pronunciation of individual sounds (potentially supported by fluency in the IPA alphabet); sense of flow and prosody. Passive competence (understanding) desirable to communicate meaning. |
| Language coaches | Native pronunciation of individual sounds, sense of flow and prosody, intimate understanding of text and how it connects to music. |
| Audiences | Not active/passive competence needed but sensibility to engage with multilingual art song programmes; use of printed translations might be encouraged (Campbell 2023). |
There are three observations I would like to make about the above regime, because they will be particularly relevant for my upcoming discussion. Firstly, readers might wonder how this linguistic regime is different from that of opera, and whether it would make more sense to talk, so to speak, about a linguistic regime of classical singing. My position is that it would certainly make sense to talk about them together (and I hope that this will be a topic of further research): for example, some of the institutions I cited in footnote 2 offer courses in ‘lyric diction’ or ‘applied languages’ which—at least if we go by their course descriptors—are meant to apply equally to opera and art song. But I would like to argue that there are a few differences, which—particularly in the case of a small-scale study like the present one—made it advisable to choose a more bounded object of study. Indeed, art song is typically thought of as a unique communion of text and music, often closely bound to national ‘high poetry’ traditions, in ways that opera is not. In Table 1, this means that some of the participants (composers, translators, coaches, perhaps singers) might, at some points, be expected to demonstrate an understanding of text at the level of versification and poetic language, which is not always expected with the comparatively simpler texts in opera.
Another difference, as anticipated in footnote 5, concerns the hierarchy of languages: Italian will certainly be seen as a core language (perhaps the core language) in the linguistic regime of opera; in art song, it has secondary importance with respect to German, English and French. This takes me to the second observation I would like to make about this linguistic regime: in this linguistic regime, different languages are given unequal importance, and these inequalities happen at the systemic rather than at the individual level. It is not normally the case that the expectations detailed in Table 1 apply only to the core languages and that lesser represented ones are treated with less rigour. A singer who has Danish as her first language and wishes to sing, say, Frederic Mompou’s songs in Catalan will likely feel the need to develop excellent pronunciation skills in this language to present his songs at the same level as she would present those by, say, Hugo Wolf or Henri Duparc. However, it is likely that she will not develop those skills unless there is a specific song or repertoire she wishes to or is paid to sing—in contrast to French, German and English pronunciation (and Italian for opera), which will have likely been a core part of her conservatoire training. Similarly, Portuguese, Nordic and Polish composers from the early 20th century were likelier to have skills in setting text in French, Italian or German than, say, the other way round. The extent to which participants engage with individual languages is therefore strongly influenced by the historical importance and prestige awarded to those languages.
My third observation is that the linguistic regime of art song is historically situated; for space and complexity reasons, this historical dimension is not necessarily captured in Table 1, but disentangling it constitutes another important avenue for research into multilingualism in art song. For example, in order to understand the linguistic expectations placed on singers, and examine them critically, as this article intends to do later on, it is important to understand how these expectations came about, and which shifts and ideologies they might have been influenced by. We know that in the late 19th and early 20th century, singing art song (and opera) in translation was more common as a practice than it is now. This did not necessarily mean, however, that singing in one’s native language was the norm, or even an option sometimes: some languages operated as musical lingua francas, so to speak, and lyrics would be translated into these with the aim of making them more accessible to a wider range of singers (and therefore audiences). For example, in London, German Lied was sung in English translation, but also in French (Tunbridge 2013: 59), whereas, in Spain, French and even Spanish opera were routinely performed in Italian (Young 2016: 128). If anything, the perceived language shortcomings of singers at this time tended to be focalised around their native languages, and not their foreign languages: when composer Tomás Bretón took up the directorship of the Conservatoire of Madrid in 1901, one of his main concerns was to make sure that student singers learned to sing in Spanish (as the teaching had been dominated by Italian singers throughout the 19th century) (Young 2016: 128). Gabriel Fauré was faced with a similar concern when becoming head of the Conservatoire of Paris four years later (in this case, the competitor was German Lied) (Bergeron 2010: 3–4).
The process by which singing art song in its original language became the expectation on stage (as is the norm nowadays; see Campbell 2023: 7) needs further examination. Certainly, it did not happen overnight: in the 1950s, the United States was still home to vivid debates on whether art song should be sung in translation or in the original, with advocates of the former, such as composer Virgil Thomson, arguing that translation enabled comprehension by the audience and singers (De’Ath 2019: 48). But this should be understood as the tail end of a process, and not as a beginning. In the absence of a more comprehensive history, I would like to suggest that there were two shifts in the interwar years which contributed crucially to the configuration of a new linguistic regime that has endured to the present day. Firstly, it was an era in which the concept of national identity, as well as individual national identities themselves, were magnified, challenged, sometimes torn apart and then reconstructed; these negotiations often expressed themselves in how art song was presented to audiences and thought about, both in their own countries of origin and abroad (see, for example, Bergeron (2010) and Tunbridge (2019) for the examples of French mélodie and German Lied). With language typically being regarded as a fundamental part of national identity, it is likely that such developments played a key role into the ‘division of the song repertoire into discrete linguistic and national groups’, which is typical of the art song ecosystem today (Bullock & Tunbridge 2021: 2). Secondly, during the interwar years, and owing to the development of electric recording, not only did record-listening became mainstream; the figure of the ‘record-listener’ also emerged during this time, meaning someone who listens to recordings (typically of Western art music) systematically, carefully and in a discerning way (Maisonneuve 2007). The enhanced listening possibilities opened up by electric recording as well as the ease of comparison between different versions revived the Werktreue ideology of the 19th century:6 it was now easier than ever to judge a performance of a given work on its degree of Werktreue, and it was also easier than ever to enjoy only those (recorded) performances showing an acceptable degree of Werktreue, rather than relying on what was on offer in the local concert hall. It is likely that the transition to a listening regime dominated by recorded rather than live listening changed standards in terms of how acceptable it was for a performance to deviate from the original score (including in terms of language).
We know a bit more about how the new linguistic regime imposed itself on educational institutions. It is likely that such changes to curricula not only reflected demand (that is, demand for art song sung in its original language) from audiences or concert programmers, but in turn contributed to further stimulating and shaping this demand. Throughout the 1960s and 1970s, use of IPA (the International Phonetic Alphabet, which allows every language to be rendered as a series of phonetic signs) became normalised in singer training, at least in the USA (Montgomery 2020: 301). Anthologies of, say, French or German art song would often feature IPA transliterations alongside the scores, texts and translations of the songs, allowing singers to figure out how to normatively pronounce words even if they did not have the support of a native coach. Another key innovation in these decades was the development of the concept of ‘lyric diction’, that is, a particular approach to diction that accommodates the principles of classical operatic technique rather than attempting to faithfully replicate oral speech. Illustrating the asymmetries in competence I discussed earlier, it is significant that attention to language teaching and learning in specialised journals in the singing profession has focused overwhelmingly on diction and pronunciation, with comprehension of the language being awarded a much less prominent role (Montgomery 2018; De’Ath 2019; Leigh 2019). Empirical studies of student singers’ aspirations have suggested that singers are indeed interested in developing reading comprehension (other than pronunciation) to understand the plots of what they are singing, but at the same time they will award less importance to traditional aspects of language-learning such as grammar (Poçi & Cilka 2015: 58–9).
Singing multilingually: solution or problem?
The last few years have seen a number of challenges to the prevailing linguistic regime of art song. This has mostly taken the form of an expansion of the resources available for singers to engage in lesser represented languages. For example, in the last few decades, anthologies of art song have been published for Japanese (Shimizu & Moteki 2014, 2016), Romanian (Boire 2004), Danish (Thaysen et al. 2024), Brazilian Portuguese (Mello & McDavit 2021), Polish (LeVoir 2025) and more languages. Digital tools have also been enlisted in pursuit of diversification, for example at the Korean Art Song Resource,7 the Ukrainian Art Song Project 8 and the Czech Art Song Initiative.9 Such initiatives might have the general aim of promoting rich but underrepresented traditions of art song, or contributing to the diversification/decolonisation of the art song stage, or both. While these are all extremely valuable initiatives in terms of expanding multilingualism in the art song ecosystem, they do so mostly in the form of expanding the palette of languages to which singers (and, by extension, their audiences) have access; however, in other respects, they adhere to the linguistic regime I have sketched out above. For example, the anthologies still rely on normative IPA transliterations as the golden standard, presented as the way to render text in a foreign language in an ‘accurate’ or ‘authentic’ way. The act of anthologising itself, no matter how carefully done (several of the anthologies above include songs by women composers, for example), risks being problematic itself, opening up hierarchies of value in the art song ecosystem to reflect a broader palette of languages, but establishing in turn hierarchies within these languages (online databases, on the other hand, by virtue of being more expansive and updateable, can get round this issue more easily). Similarly, there is little understanding of how singers and singing students might engage with these resources while still responding to other pressures typical of the art song ecosystem: for example, if a conservatoire student is required to take repertoire classes in French, German and English (and Italian for opera), how likely would she be to add another language into the mix, particularly if a repertoire coach is not readily available? And might an early-career working singer feel that building skills in a lesser represented language takes away from her perfecting major arias and songs in her voice type that audiences and audition panels will expect to hear?
While not purporting to offer a solution to ‘the problem’, the rest of the present article draws upon practice-led work and qualitative research with singers to articulate how singer engagement with songs in lesser represented languages happens within the linguistic regime described above, and how it might be particularly helpful in terms of teasing out and critically analysing some components of this regime. My argument is that, rather than simply asking ourselves how singers can be supported in learning repertoires in lesser represented languages, we need to consider that the (perceived) barriers to doing so might not always be personal or institutional shortcomings, but rather are key features of a linguistic regime which, in encouraging certain languages and multilingual practices, stifles others.
Methodology
The data I discuss in the second half of this article result from an experiment designed to gain insights into the strategies that singers develop when learning art song in languages unfamiliar to them; the experiment’s null hypothesis was that there is no single strategy that all singers use in such situation, but instead these strategies are highly adaptative, and can even be innovative and creative.
For the experiment, after ethical approval was obtained from my institution, I recruited four singers through a call for participants disseminated at my university’s Music department and a local conservatoire. The fifth participant was myself. While all singers are now based in the same city, their educational backgrounds (as reflected in Table 2) include six different countries. Singers were asked to select an art song in a language they were unfamiliar with (that is, a language they did not speak, they had never studied, and they had never sung in)10. In order to facilitate choice, participants were given access to a directory of art songs in about twenty languages that resulted from the same funded project that the present article comes from. It was made clear to participants that this directory was meant to serve as inspiration only, and participants were free to choose other songs if they wished. However, in the end, all singers chose songs included in the directory. Furthermore, in order to encourage choice on artistic rather than purely pragmatic grounds, it was made clear to participants that the project’s administrator could procure/acquire scores in case these were not freely available online.
Table 2
Singers’ background and chosen song.
A table that gives details about the 5 singers who took part in the project, incl. native language and language and name of chosen song.
| Singer ID | Status | First language | Language of song | Chosen song and composer |
|---|---|---|---|---|
| S1 | Unergraduate student | English | Welsh | ‘Mai’, Meirion Williams |
| S2 | Postgraduate student, semi-professional singer | Mandarin Chinese | Greek | ‘Tin eides tin xanthoula’, Spyridon Xyndas |
| S3 | Semi-professional singer | English | Japanese | ‘Sakura Yokochō’, Sadau Bekku |
| S4 | Undergraduate student | English | Mandinka | ‘Duniya’, Fred Onovwerosuoke |
| S5 | Semi-professional singer | Spanish | Polish | ‘Sroczka’, Grażyna Bacewicz |
Participants were given three months to learn their chosen song, in preparation towards a workshop open to the public. I interviewed them a month into the learning process, and then again three months into the learning process (approximately one week before the workshop took place) (questions for all interviews can be found in the Appendix ). In my own practice-led process of learning my own song, I replaced interviews with a journal, which I filled in after each learning session with the song (which happened at irregular intervals). Journal entries did not follow a predetermined format, to account for differences across the learning process; however, I regularly revisited the questions I was asking to other singers and strived to cover the same ground in my entries (text, pronunciation, meaning, music, etc).
At the workshop, which gathered an audience of about thirty university staff and students as well as external musicians, each singer was allocated a half-hour slot to work with a language coach. For sustainability reasons, language coaches were recruited predominantly from the local music scene. Because of the range of languages represented, it became obvious early on that the definition of ‘coach’ needed to be more flexible than is usually understood in the art song ecosystem: two of the coaches are professional musicians (a pianist and a singer) highly proficient in the art song tradition, who also coach professionally in their native languages upon request. In cases in which it was not possible to obtain this specific skillset, the priority became to recruit individuals who were first-language speakers of the language in question and had professional expertise both in languages (for example, as a language teacher, or as a researcher in Linguistics) and in music (not necessarily as a musician, but as a producer). All coaches were briefed on the context and purpose of the workshop, and they were advised to treat the coaching session as a co-created process, in which both the coach and the singer were welcome to put forward their priorities, requests and needs in a supportive way. Within one week of the workshop taking place, singers were interviewed for the third time.
The decision to design an experiment—that is, an intervention in the participants’ creative lives—rather than, say, conduct interviews or a survey with a larger number of singers with previous experience of learning songs in unfamiliar languages had to do with a desire to obtaining greater granularity and alignment. Designing an experiment allowed me to follow singers in real time as they learned their songs, prompting reflection during and shortly after the fact, whereas asking participants to refer to past experiences might have run the risk of reflections being less vivid. The same can be said about the numbers of participants: while I am aware that a greater number of participants would likely have resulted in a greater range of strategies, a smaller number of participants, including my own practice-led research, allowed me to carefully track each singer’s reflection process.
I was aware, on the other hand, that the experiment I designed risked being somewhat artificial and not reflective of working conditions. I would argue, however, that in real life there is no single set of ‘working conditions’ associated with learning a song in an unfamiliar language: some singers might do so, for example, in response to a request from a particular concert promoter (say, a singer who is asked to perform in front of a culturally specific audience), which might remove the element of choice, but will likely introduce a financial and/or professional incentive, as well as pressure to learn the song and master pronunciation by a certain deadline. There might be other singers who learn repertoire in an unfamiliar language on their own initiative, over the course of several months or years, not for the purpose of singing it at a specific date, but rather with the aim of including it in their repertoire once they feel sufficiently comfortable. As a compromise, my experimental design still included two features which are typical of the working lives of singers and singing students, and which tend to shape heavily repertoire learning processes: firstly, the experiment included a workshop open to the public, which, while not a performance, required that the singers felt sufficiently confident presenting their work in public by a certain date; secondly, it also included half an hour with a language coach, which is a feature of conservatoire education and the working lives of some singers.
After the experiment concluded, I coded interview transcripts and my own journal using a bottom-up coding framework, that is, letting the codes emerge from the interviews and journal themselves. In the next section, I group codes into different categories, with an eye on how these groupings helps us shed light on and problematise aspects of the linguistic regime of art song and what happens when lesser represented languages are included within it.
Findings
Choosing a song: the music itself?
(codes: melody; text–music connection; composer)
When choosing a song, perceived musical quality played a role for all participants. Extra-musical considerations played a role for two of the singers, but never in isolation: S4 started the process by wanting to sing in a specific African language because of a pre-existing family connection; however, after being unable to find a song in said language which they liked/felt comfortable singing, they moved on to selecting another more musically appealing song in another African language, which also had a personal meaning for them due to a social connection. S5 initially selected their song because of its ‘liveliness’, but a further factor that played a role was having a partner who was also a speaker of the language the song was in.
A common reason for choosing a song was its melody (S1: ‘[when I first heard the song] I felt I could really get into the melody’; S2: ‘[The melody] is so happy’; S3: ‘I like songs that move from low to high registers’). Apart from this initial interest, some of the participants also reported that, as the learning process progressed, they started to think more deeply about the relationship between text and music in their particular song and/or more broadly in the language in which they were singing. For example, throughout interviews 1 and 2, S3 spoke about a ‘unique fit’ between the text and music, which they felt differed from what they were used to in core art song languages. S4 reported a similar feeling (‘The feel of the song is different from any song I’ve done before.’) S5 felt they were able to get a greater sense of text–music connection after receiving support from the language coach in fixing pronunciation issues (‘Fixing the pronunciation made the text have much more sense together with the music.’) S1 reported that the most useful takeaway from witnessing the other singers’ coaching sessions was to see ‘how each language works with the music’. These two last responses are consonant with the linguistic regime described above, in that they confirm the importance that is awarded to coaches and the process of coaching in successfully singing in an unfamiliar language.
From word to phrase to meaning
(codes: pronunciation; text; phrasing; meaning; language families)
Predictably, a challenge that most singers encountered later on was that of learning the text to their chosen song. Participants were not given any specific guidance or resources about how to do this, as one of the aims of the experiment was to determine which strategies singers came up with themselves; they were not prevented from seeking online tools (which will be discussed in the next section) or asking for someone’s help (for example, a friend or family member who speaks the language), as it was acknowledged that singers might well use such strategies in a real-life context. This means that they were not prohibited (but similarly not explicitly encouraged) from deviating from the learning model usually prescribed in the linguistic regime, in which work with a coach and IPA transliterations are often preferred to other learning methods. Most participating singers chose to separate the processes of learning the music and learning the text, starting with the former—sung to ‘la’ or another syllabic sound—and only subsequently introducing the latter (S1, S2, S3). All three singers noted that this was different to their usual learning process in a familiar language, where they would have learned the words and music together from the beginning. S5 also departed from their usual process of learning music, albeit in a different way:
I had planned to start the process by getting an overall sense of the song, getting to know the melody and shape well before focusing on individual phrases or words—this is the process I usually follow. However I found myself unable to move past the first phrase, wanting to nail down the details before moving on.
In learning the sounds, all singers resorted to recordings of the song (both audio and video), from which they tried to copy the pronunciation, and two also used IPA transliterations (in one case made by the singer themself from a video, which, as I will discuss in the next section, was not unproblematic; in another, kindly facilitated by a native speaker of the language). S3 worked with an anthology which included a general pronunciation guide of their chosen language. Other online resources used by the singers—which deviate more from what is usually prescribed in the predominant linguistic regime—are discussed in the next section. While most singers reported encountering some difficulties with this process, they also found it rewarding: S1 described it as ‘quite interesting’ to learn about the different sounds in their chosen language. Most of the singers reported almost instinctively leaning on their knowledge of other languages to understand pronunciation: Italian for S2 (while singing in Greek), Russian for S5 (while singing in Polish). For other singers, while not drawing on similarities to the same extent, other languages acted as a point of reference: for example, S3 reported feeling that vowels are ‘much softer’ in Japanese than they are in German. S4 was the only participant who reported attempting to draw upon other languages, but not finding it productive (‘that was not very useful here’).
A challenge that most singers noted was moving from pronunciation of individual sounds to what they variously called ‘prosody’, ‘flow’, ‘phrasing’ or ‘connection’; while these terms are not perfect synonyms, in this context they are taken to refer to how a phrase or sentence (as opposed to a single word) sounds in the original language: whether there are elisions or pauses, for example, or whether some words or syllables will feel more present than others. Singers tried to obtain this knowledge, again, from recordings (S2, S5) and, in one case, from spoken recordings of the language procured online (S4). Three of the singers reported that this led them to ask themselves how the grammar of their chosen language works (S2, S3, S5), in an attempt to understand broader peculiarities of prosody and flow—for example, understanding where the verb typically sits in a sentence (S3, S5). Singers felt that obtaining this knowledge could be important to informing musical decisions: for example, in terms of where to breathe (S2, S5). Two of the coaches (for S1 and S3) chose to offer during their sessions a short introduction on how the language ‘sounds’ with respect to English (for example, whether consonants are prominent with respect to vowels, or vice versa, or whether they are more discreetly fitted into the flow of the speech), which both singers found useful to help them with the sense of flow and prosody.
With much of the focus both during the learning process and subsequently during the coaching sessions being on pronunciation and, secondarily, flow, most singers felt that they could not work on issues of meaning and expression as much as they would have liked (which typically play a role in the learning process prescribed by the predominant linguistic regime). Generally speaking, before their coaching session, singers aimed to get the general sense of the text of their song, but not a detailed understanding of every word or sentence, and none of them used notes or annotations systematically to work on meaning. S2, for example, reported that it was important for them to understand the ‘composers’ emotion’; S4 said that realising their song was meant to be a prayer gave her a perspective on ‘how to hold myself’. The extent to which coaches were able to work on meaning during the session varied; this was easier for the shorter songs (S4, S5), for which coaches explained the meaning of each word as well as aspects of register, grammar and tone in each sentence, then asking the singer to tailor their performance to convey that. When there was less time to work on each sentence individually (S1, S2), coaches still gave a general summary of the text (which included not only content, but also discussion of some elements of tone, register and context), and invited the singer to give a final performance while keeping this in mind.
Use of online tools
(codes: online tools; translation; in-person; native speaker)
To a greater or lesser extent, all singers made use of online resources during their learning process. While the peculiarities of online tools, with their potential for constant updating and expanding as well as multimedia content, have been used to a moderate extent in supporting language skills for singers,11 these are still rather localised developments, and the focus remains on more traditional tools, such as anthologies, as above (perhaps partly because of the greater convenience afforded by learning and annotating songs on paper scores). However, with generic online tools having an increased presence in the lives of most of us, including singers, it is predictable that these will be increasingly used throughout the learning process—deviating from what has so far been the norm within the predominant linguistic regime. Singers reported using: Google Translate (both for obtaining translations of the song’s text, and to aid with pronunciation), DeepL (for translations), ChatGPT (for translations), YouTube (for both recordings of their chosen song and for spoken videos in their chosen language), translations and contextual information published online, and online pronunciation guides.
While being able to access these resources initially engendered a certain sense of optimism in some of the singers, the process was not straightforward. Whereas two singers were able to easily find multiple recordings of their chosen song on YouTube (S1, S5), other singers could find only one recording of their song (S2, S3, S4); two of these (S3, S4) also noticed that the singer’s performance contained discrepancies from the printed music, which led them to wonder whether the sung text might contain similar inaccuracies. S1 reported trying three different tools to translate their song, with discrepancies and differences being present between the three versions. In other cases, the shortcomings of online tools became more apparent during the workshop, as singers found that their attempts at copying sounds from a recording might not always match standard pronunciation in the language (S2: ‘On YouTube the sound is so fast, you cannot hear each sound clearly.’)
After the workshop, all singers highlighted the advantages of being able to collaborate face-to-face with a native speaker (‘The tools online cannot compare to talking to a real person. I learned so much more from [coach] than online’ (S1)), although some (S1, S2) acknowledged that they would have needed more time with a coach to work more comprehensively on their song. Still, some singers emphasised the need to be pragmatic when learning a song in a foreign language, as face-to-face time with a native speaker might not always be an option (S4: ‘I might have to consult YouTube if I cannot find anybody’; S3: ‘I would be interested in learning songs in other lesser represented languages, but it depends on what is available.’)
Ethical considerations
(codes: ‘Western-centric’; ‘perfection’; ‘respect’)
In the context of decolonial approaches to music education, researchers have warned about the risk of adopting a ‘touristic’ model that engages with different musical cultures only superficially and for the purposes of showcasing surface diversity (Wasiak 2009; Hess 2015; Sánchez-Gatt et al. 2025). There is a risk that similar concerns apply to the linguistic diversification of the art song language regime. These might take different forms, some of which are more prevalent in the present-day art song ecosystem, while others are more theoretical at present. For example, while songs in hegemonic languages will be routinely included in recital programmes based solely on their perceived quality and/or historical significance, songs in lesser represented languages typically need a further reason to be taken into consideration (for example, being in the singer’s native language, or being part of a language- or country-themed event). Sometimes, engagement either with the language itself or with the art song tradition in a given language might be superficial, with songs in lesser represented languages being regarded or presented as derivative or imperfect with respect to hegemonic traditions. Sometimes, singers, teachers and programmers dedicate efforts to expanding and diversifying repertoire in the major languages (for example, by including songs by women composers), but might consistently go back to the same choices when it comes to including a lesser represented language (for example, Granados’ Tonadillas, Falla’s Siete canciones populares españolas and Rodrigo’s Cuatro madrigals amatorios for Spanish art song).
At the learning stage, such concerns did not feature heavily; only one of the singers (S4) expressed concern at singing a song in an African language while ‘being very white and Scottish. I try not to get this into my head too much.’ However, they emerged more clearly during and after the workshop. There were two key developments in the workshop that seemed to stimulate much of the reflection. Firstly, singers were able to acquire—both from their own coaching sessions and from other singers—some sociolinguistic knowledge that challenged their understanding of concepts such as ‘norm’, ‘standard’ and ‘dialect’. For example, they learned that Mandinka comprises several dialects and variants, and so the question of what is the ‘correct’ or ‘standard’ pronunciation of a given word in the song could vary depending on the speaker answering the question (S4: ‘it struck me that there are different ways of interpreting the text’). They also highlighted learning that speakers of North and South Welsh, pre-standardisation, would not have always understood each other (S4), and that the text of the Japanese-language song contained several archaic phrasings, meaning that it would not necessarily have sound natural to a contemporary native Japanese speaker (S3). Secondly, most singers found that they needed to spend more time learning pronunciation before they could move on to focusing on expression and meaning (S4: ‘It’s tough to learn some consonants if you haven’t learned them in the first years of childhood.’)
These developments during the workshop prompted some of the singers to reflect critically on notions of perfection and accuracy in pronunciation, and particularly on how these are typically centred in the expectations imposed on singers in the linguistic regime described above. Taking the Mandinka song as a cue, singers reflected on how notions of perfection might not even apply to all language coaching situations—for example, in the case of languages with no universally accepted standard and no or little art song tradition (and of how the language is pronounced within it)—and they articulated how such notions are typically more readily applicable to the dominant languages:
In speech we are so used to English everywhere, but in art song when something is in English there’s that one standard, RP sound. You get a lot of the rhotacity of Scots beaten out of you when learning to sing. (S4)
[Some teachers or coaches emphasize] how important is to get every word right but that sounds a bit of a Western-centric approach but there are other languages where this is not as important. (S3)
Singers, however, also rejected the notion that pronunciation and accuracy are completely unimportant, and articulated the view that singers still have certain responsibilities and expectations with respect to text. S3 summarised it as:
respect and effort […]. Sometimes it’s easy to be a lazy singer and just sing notes but there’s that extra effort related to both the music and languages that we all have responsibility to engage with.
S2 emphasised that having access to a language coach ‘make[s] the work with the language feel more realistic’, and advocated for a pragmatic approach where work with language (including pronunciation, meaning and context) focuses on what they regarded as the aim of a performance:
We should still try to make the words as clear and meaningful as we can and […] sentences—as clear as we can and honest in our expressions. We don’t have to be perfect.
Conclusions
From my analysis conducted under the headings above, I would like to offer some final reflections about how the findings from this experiment can open up directions to encourage and support further engagement with lesser represented languages, as well as (more ambitiously) to rethink the linguistic regime of art song. While the sample was limited due to the time- and resource-intensiveness of the experiment, I believe that the relative diversity of the sample12 has served to illuminate a range of issues that are likely to recur, in more granular form, in further research with larger numbers of participants. First of all, the experiment confirmed how time- and resource-intensive this regime can be for singers: to learn a new song in an already familiar, ‘standard’ language (French, German, English) typically requires considerable investment from the singer, both in terms of time and money. In the case of an unfamiliar language, this investment is even greater, because the singer does not even have a foundation to build upon (for example, previous experience of singing in the same language, access to a dedicated class or coach if they are studying at an educational institution). This, indeed, seems a serious obstacle for singers to engage with lesser represented languages more consistently. In response to this, I see two possible directions here which are not necessarily fully incompatible, but in fact might be applied in different contexts or at different times: the first of those would be to accommodate lesser represented languages within the existing linguistic regime, and the second would be to challenge more decisively some aspects of this linguistic regime.
I will start by discussing the first of these options: that is, expanding or tweaking the regime slightly to include lesser represented languages. Here, I would like to make three points. Firstly, as discussed above, all singers in the sample attempted—with varying levels of success, depending on resources available and previous experience—to move from pronunciation of isolated sounds, to ‘flow’, to meaning, figuring out in the process how these different components might be connected in the context of their chosen song and language (for example, how the word order or conjugation system typical of the language can affect flow, influencing both pronunciation and meaning). While such connections of course happen in more mainstream languages too, some participants felt that the experience of engaging with a less familiar language made them more aware of these (S2, S3, S5). A potential path forward when supporting singers in engaging with lesser represented languages might therefore involve moving beyond the usual focus on pronunciation of individual sounds and more towards matters that can support musical decisions like phrasing and expression: for example, details about prosody, or selected details about grammar. It would, of course, be unrealistic to provide a full ‘primer’ to a language, particularly since there is an element of unpredictability in terms of which elements of language a singer might find productive to inform her performance. A solution could be to enlist the experiences of teachers and coaches, as well as those of singers who have attempted to sing in lesser represented languages, in order to collaboratively design resources that might draw attention to aspects of prosody, flow and grammar that differ from the major languages and can hence inform performance decisions more decisively.
Secondly, results from the experiment also suggest that enlisting familiarity might be a productive strategy for singers to engage with the process above: even if a singer is not familiar with a given language, they will be familiar with the process of singing in a foreign language, and so it is perhaps inevitable that they will rely on their previous knowledge on sounds or structures they know from other languages. While this can certainly lead to error (for example, even languages in the same language family will not share the same set of phonemes or grammar features), it can also make languages more accessible. As above, systematising this might entail relying on the expertise of coaches, teachers and singers to build resources that identify sounds or phrasings in German, French or English that might help singers with aspects of other languages. This is likely to be of greater help for closely related languages (for example, German for other Germanic languages, Italian for Spanish), but the example of the Mandarin-native singer (S2) relating their Greek text to their knowledge of Italian suggests also that distance can be relative. There is scope, I believe, to make use of such relative proximity to help singers understand how words and sentences are built in an Indo-European versus a Sinitic language (or indeed potentially other language families).
Thirdly, as younger generations of singers make extensive use of online tools and apps in various aspects of their lives, there is scope to think about how these aids can be engaged productively to encourage singing in lesser represented languages. The results suggest that caution is needed when using generic tools (intended for communicative contexts typical of contemporary everyday life) to obtain pronunciation and interpretation support regarding highly complex poetic texts intended for performance. Although this did not come up in interviews, the ethical and environmental issues of using Large Language Models should also be taken into account. But, curated online tools are worth exploring in combination with the points made in the section above: for example, overviews of prosody and ‘flow’ could be supplemented with YouTube videos that offer practical examples of the spoken language, and online tools can also be more regularly updated or even crowdsourced to welcome perspectives from a range of singers and coaches.
Other takeaways from the experiment, however, seem more difficult to address without more decisively challenging the linguistic regime of art song (I note, for example, that some of the strategies I have discussed so far would likely contribute to further centring already hegemonic languages: for example, by ‘relating’ other languages to them). Here, I would also like to make three points. Firstly, as I have hinted at above, resources seeking to introduce a greater array of languages into art song tend to be themselves monolingual, reflecting only one language and replicating the language/national units into which art song is typically organised. While this certainly ensures depth and specialisation (and might also be more easily legible in the current art song ecosystem), it also raises questions as to what the pathway might be through which singers discover these resources: the current system seems to presuppose that language choice comes first and song choice comes second, and that a singer will start by developing sufficient interest in, say, Korean or Japanese art song first to specifically search a directory or acquire an anthology. However, the results from the experiment show that this might not always be the case, and that more open-ended possibilities for the discovery of underrepresented repertoire that are not bounded by language might be more encouraging for some singers.
Secondly, the extent to which singers mentioned the text–music connection as being important to them and, in some cases, surprising, also opens up avenues for rethinking some aspects of the linguistic regime. I would like to acknowledge that the singers’ comments on the uniqueness of text-setting in their chosen languages is not based upon close, sustained study of said language and so they must not be taken at face value; however, I argue that such comments can be revealing in a different way. The intimate connection between text and music (more than merely ‘setting words to music’) is seen as a fundamental pillar of the art song genre, and musicologists, teachers and singers themselves have dedicated and continue to dedicate considerable energy to disentangling how such a connection expresses itself in songs by canonic composers, and more generally in the major languages. In this context, engagement with lesser-known languages can instil a sense of curiosity and discovery, not simply at the language itself, but also at how it works in combination with music—particularly, for example, in the case of Mandinka and Japanese, which might not be able to rely in music on the syntactic structures which work for the Indo-European languages. This can in turn encourage a less Euro-centric approach to the genre, understanding songs in lesser represented languages not as mere substitutes for songs in hegemonic languages which draw upon the same techniques and musical styles, but rather as unique creative products which embody different, language-mediated ways of approaching the connection between text and music.
Thirdly, engagement with lesser represented languages can encourage critical reflection on notions of ‘perfection’, ‘accuracy’ or ‘standard’. Inclusion of lesser represented languages into the repertoire might seem to inevitably lead to a trade-off: focus on perfection and complete accuracy might make singers less likely to engage with lesser represented languages due to the investment in time (and, in some cases, money) required; at the same time, relaxing such standards to make lesser represented languages more accessible could result in a ‘touristic’ approach to repertoire in which the depth of engagement expected for a song in German might not apply to, say, a song in Finnish or Arabic. However, more systematic inclusion of lesser represented languages in teaching and the repertoire could in fact help singers, coaches and teachers negotiate both extremes in practice: to figure out what ‘respect’ and ‘effort’ might mean in specific contexts, and to figure out just how much linguistic input is needed to allow the singer to deliver an expressive, communicative performance. There might be broader benefits from engaging with lesser represented languages: as suggested by the Mandinka example, such languages can help students develop a critical perspective on issues of ‘norm’, ‘standard’ and so on, which is not only relevant for individual lesser represented languages, but also for their approach to languages as singers more generally: in the 19th century (an important repertoire for most singers), the development of art song in French and German was intrinsically connected to the standardisation and codification of these three languages as national languages.
Appendix
Sample questions for interview 1 (one month into the learning process)
Can you briefly describe your experience singing art song/opera? In particular, which languages have you sung in in the past? If these are not native languages, do you speak those languages, and to which level of fluency?
What were the reasons why you chose this song?
Can you describe your first steps into the learning process? E.g. did you sight-read through the song? Listen to a recording? Focused on a specific passage or try to get a glimpse of the whole song?
Can you describe your first steps specifically relating to text? E.g. did you try to find a translation of the text? An IPA transcription? Did you listen to a recording of the song sung by a native speaker? Did you maybe not concern yourself with text at all?
Did your attitude to text change throughout the learning process? Did you seek any of the resources mentioned in the question above, or others?
How has it felt so far to learn a song in a language you were not familiar with? Has it felt different to singing songs in your native language? In a language you have experience of singing in?
How would you say comprehension and expression relate for you at this stage? E.g. are you trying to understand every word and give it a distinct expressive quality? Or are you focusing more on the ‘general’ feeling or mood conveyed by the song? Or something in between? How does the music factor in in this?
Sample questions for interview 2 (three months into the learning process)
Can you describe your learning process in the last few months? What would you say have been your major breakthroughs with this song?
Did your attitude to text change throughout the learning process? Did you seek any of the resources mentioned in first interview (IPA transcriptions, translations, etc.), or others?
How has it felt so far to learn a song in a language you were not familiar with? Has it felt different to singing songs in your native language? In a language you have experience of singing in?
How would you say comprehension and expression relate for you at this stage, and has this changed since interview 1? E.g. are you trying to understand every word and give it a distinct expressive quality? Or are you focusing more on the ‘general’ feeling or mood conveyed by the song? Or something in between? How does the music factor in in this?
Sample questions for interview 3 (after the workshop)
What was the most useful thing you learned from your session at the workshop? What was not so useful?
What was the most useful thing you learned from other participants’ sessions at the workshop? What was not so useful?
What was the most useful thing you learned from the final discussion at the workshop? What was not so useful?
Did you find your coach’s approach useful/unexpected/predictable, in regards to focus (i.e. did they focus predominantly on pronunciation of individual words? General expression? Something in between? Something else?). How useful do you think this approach might be for singers not singing in their native language?
In the future, do you think you will sing more songs in [language of the song you chose to sing]? If the answer is yes, what kind of help or support do you think you might need to achieve this?
In the future, do you think you will sing more songs in other languages you don’t speak? Which ones? If the answer is yes, what kind of help or support do you think you might need to achieve this?
References
Ayres-BennettW. & FischerL. (2022), ‘Towards interdisciplinarity in multilingual identity research: differing perspectives and common ground’, in
Ayres-BennettW. & FischerL. (eds), Multilingualism and Identity: Interdisciplinary Perspectives (Cambridge, Cambridge University Press), 1–18. https://doi.org/10.1017/9781108780469.001
BergeronK. (2010), Voice Lessons: French Mélodie in the Belle Epoque (New York, Oxford University Press). https://doi.org/10.1093/acprof:oso/9780195337051.001.0001
BinderB. (2021), ‘Performance matters in heine: the case of Pauline Viardot’s “Das ist ein schlechtes Wetter”’, in
BullockP. R. & TunbridgeL. (eds), Song Beyond the Nation: Translation, Transnationalism, Performance (New York, British Academy/Oxford University Press), 94–113. https://doi.org/10.5871/bacad/9780197267196.003.0006
BoireP. (ed.) (2004), Romanian Art Songs. Soprano (Webster, NY, Leyerle).
BullockP. R. & TunbridgeL. (eds) (2021), Song Beyond the Nation: Translation, Transnationalism, Performance (Oxford, British Academy/Oxford University Press). https://doi.org/10.5871/bacad/9780197267196.001.0001
CampbellS. (2023), ‘“Oh for heaven’s sake, do I need to explain this really?” translation Skopoi in live art song concerts’, Translation Review, 116(1): 1–12. https://doi.org/10.1080/07374836.2023.2231039
De’AthL. (2019), ‘Early articles on language and diction in NATS journals’, Journal of Singing, 76(1): 47–54.
De FrancisciE. & MarinettiC. (eds) (2025), Translation in the Performing Arts: Embodiment, Materiality and Inclusion (Abingdon, Routledge). https://doi.org/10.4324/9781003335573
FootittH. (2022), ‘Multi-lingualism’, in
BurnsJ. & DuncanD. (eds), Transnational Modern Languages (Liverpool, Liverpool University Press), 217–225. https://doi.org/10.2307/j.ctv2fjwpw7.28
GamratM. (2023), Franz Liszt’s Songs for Voice and Piano (Leiden, Brill). https://doi.org/10.1163/9789004548862
HessJ. (2015), ‘Decolonizing music education: Moving beyond tokenism’, International Journal of Music Education, 33(3): 336–347. https://doi.org/10.1177/0255761415581283
KohlK. & OuyangW. (2020), ‘Introducing creative multilingualism’, in
KohlK., DudrahR., GoslerA., GrahamS., MaidenM., OuyangW. & ReynoldsM. (eds), Creative Multilingualism: A Manifesto (Cambridge, Open Book Publishers), 1–25. https://doi.org/10.11647/OBP.0206.12
LeighS. A. (2019), ‘The seven point circle and the twelve principles: an evidence-based approach to italian lyric diction instruction’, Scenario, 2: 184–198. https://doi.org/10.33178/scenario.13.2.12
LeVoirL. (2025), Polish Song Anthology for Medium-High Voice (Unknown, Self-published).
LewisG., JonesB. & BakerC. (2012), ‘Translanguaging: origins and development from school to street and beyond’, Educational Research and Evaluation, 18(7): 641–654. https://doi.org/10.1080/13803611.2012.718488
Lo BiancoJ. & AroninL. (2000), ‘Introduction: the dominant language constellations: a new perspective on multilingualism’, in
Lo BiancoJ. & AroninL. (eds), Dominant Language Constellations (Cham, Springer), 1–18. https://doi.org/10.1007/978-3-030-52336-7_1
LogesN. & TunbridgeL. (eds) (2020), German Song Onstage. Lieder Performance in the Nineteenth and Early Twentieth Centuries (Bloomington, Indiana University Press). https://doi.org/10.2307/j.ctv104t9zt
LowP. (2006), ‘Translations of songs: what kind matches your purpose?’, Journal of Singing, 62(5): 505–513.
MaisonneuveS. (2007), ‘La Voix de son Maître: entre corps et technique, l’avènement d’une écoute musicale nouvelle au XXe siècle’, Communications, 81: 47–59. https://doi.org/10.3406/comm.2007.2457
MarinettiC. & De FrancisciE. (2022), ‘Introduction: translation and performance cultures’, Translation Studies, 15(3): 247–257. https://doi.org/10.1080/14781700.2022.2126386
MateoM. (2013), ‘Multilingualism in opera production, reception and translation’, Linguistica Antverpiensia, 13: 326–354. https://doi.org/10.52034/lanstts.v0i13.58
MelloF. & McDavitC. (eds) (2021), Brazilian Art Song Anthology (Barcelona, Mundo Arts).
McKinneyC., MakoeP. & ZavalaV. (2023), ‘Introduction: Critical and decolonial approaches to multilingualism in global perspective’, in
McKinneyC., MakoeP. & ZavalaV. (eds), The Routledge Handbook of Multilingualism, 2nd edn (Abingdon, Routledge), xx–xlv. https://doi.org/10.4324/9781003214908
MontgomeryC. (2018), ‘The voice and diction connection: a diction instructor’s approach to voice pedagogy’, Journal of Singing, 74(3): 313–321.
MontgomeryC. (2020), ‘Diction (still) belongs in the music department’, Journal of Singing, 76(3): 301–309.
PoçiA. & CilkaI. (2015), ‘Studienbegleitender Deutschunterricht an der Universität der Künste in Tirana. Eine empirische Studie zum Unterricht für Gesangsstudierende’, Zeitschrift für Interkulturellen Fremdsprachenunterricht, 20(1): 56–68.
Sánchez-GattL., MenonS. A. & HessJ. (2025), ‘Troubling transcultural practices: anti-colonial thinking for music education’, Action, Criticism, and Theory for Music Education, 24(1): 48–82. https://doi.org/10.22176/act24.1.48
ShimizuK. & MotekiM. (eds) (2014), Japanese art song anthology. Volume 1. High Voice (Fayetteville, Classical Vocal Reprints).
ShimizuK. & MotekiM. (eds) (2016), Japanese art song anthology. Volume 2. High voice (Fayetteville, Classical Vocal Reprints).
Susam-SarajevaS. (2008), ‘Translation and music: changing perspectives, frameworks and significance’, The Translator, 14(2): 187–200. https://doi.org/10.1080/13556509.2008.10799255
ThaysenE. H., SchachtenhaufenR. & TeglbjærgC. S. (2024), Danish Diction in Art Song (Copenhagen, The Royal Danish Academy of Music).
TunbridgeL. (2013), ‘Singing translations: the politics of listening between the wars’, Representations, 123(1): 53–86. https://doi.org/10.1525/rep.2013.123.1.53
TunbridgeL. (2019), Singing in the Age of Anxiety: Lieder Performances in New York and London between the World Wars (Chicago, IL, Chicago University Press). https://doi.org/10.7208/chicago/9780226563602.001.0001
YoungC. L. (2016), Music Theater and Popular Nationalism in Spain, 1830-1930 (LA, Baton Rouge, Louisiana State University Press).
WasiakE. B. (2009), ‘Countering musical tourism and enacting social justice: Repositioning music education as a cross-cultural meeting place’, in
GouldE., CountrymanJ., MortonC. & Stewart RoseL. (eds), Exploring Social Justice: How Music Education Might Matter (Toronto, Canadian Music Educators Association), 212–224.
1
For example, London’s Wigmore Hall (which describes itself as ‘London’s iconic venue for chamber music and song recitals’) put on eight concerts in May 2026 focusing totally or partially on art song. Four of these included songs in at least two different languages (two of the ‘monolingual’ concerts focused on large-scale song cycles—Messiaen’s Harawi and Robert Schumann’s Dichterliebe—that are typically performed as single-item programmes). See https://www.wigmore-hall.org.uk/tags/song?archivedPerformancesOnly=true &contentTypes=performance&learningPageOnly=false&page=1&startDate=2026-05-01&watchPageOnly=
false (last accessed May 2026).
2
For example: Conservatoire Supérieur de Paris (German and Italian lyrical diction is compulsory in the first cycle of studies, https://www.conservatoiredeparis.fr/en/curriculum/1st-cycle-singing), Hochschule für Musik und Theater Hamburg (module on International Phonetic Alphabet pronunciation for singers, including German, French, Italian and English, https://www.hfmt-hamburg.de/studium/studienorganisation/vorlesungsverzeichnis/ws24-25-ipa-internationale-phonetik-4390), Escuela Superior de Canto, Madrid (modules on diction in German, English, French and Italian applied to singing, https://escm.es/loe/), modules in German, Italian, Russian, French and English diction (https://catalog.oberlin.edu/preview_entity.php?catoid=56&ent_oid=5168), Boston Conservatoire (song repertoire modules in Italian, French, German and English, https://bostonconservatory.berklee.edu/voice-and-opera/bm-voice-performance). Some institutions do not give details about their study plans online, but they include named language coaches among their members of staff: e.g., Royal Academy of Music (https://www.ram.ac.uk/study/departments/vocal-studies/people), Royal College of Music (https://www.rcm.ac.uk/vocal/vocalprofessors/), Guildhall School of Music (https://www.gsmd.ac.uk/study-with-guildhall/music/music-department-staff/vocal-studies-department-staff), Universität für Musik und darstellende Kunst Wien (https://www.mdw.ac.at/gesangundmusiktheater/personen-fachbereiche/lied-oratorium-konzert/) (all links were last accessed in May 2026).
3
For example: Guildhall School of Music (a ‘variety of languages and styles’, https://www.gsmd.ac.uk/study-with-guildhall/music/music-auditions-interviews/vocal-studies-auditions), Koninklikj Conservatorium Den Haag (video auditions with French and German songs for the Masters with a Lied specialisation (https://www.koncon.nl/en/programmes/masterspecialisations/vocal/masterspecialisatie-lied/entry-requirements), Conservatoire Supérieur de Paris (works in French, German and Italian for entry into the second cycle of studies, https://www.conservatoiredeparis.fr/en/curriculum/2nd-cycle-singing), Peabody Institute/Johns Hopkins University (works in French, German, Italian and English for entry into DMA degree, https://peabody.jhu.edu/academics/instruments-areas-of-study/voice/auditions/), Indiana University (art song in German, French and English required for Masters admission, https://music.indiana.edu/admissions/auditions/voice.html), Juilliard School (compulsory English, Italian and German works, and optional French, for Masters admission, https://www.juilliard.edu/arm/music/college/voice/master-music#chapter-application-materials). Audition requirements for undergraduate study might request a smaller number of languages (e.g., the language of the country the institution is in, plus one additional language, sometimes left to the candidates’ choice), whereas postgraduate requirements tend to be more specific.
4
Exceptions include Hanns Eisler’s Hollywooder Liederbuch (1943) (comprising songs in English and German) and Jean Sibelius’s Seven songs, Op. 7 (1904) (comprising two songs in Finnish and five in Swedish). While not exactly obscure, neither of the cycles can be regarded as a popular choice in art song recital programmes.
5
I refer to footnotes 2 and 3 above—institutions which name specific languages as part of their audition requirements or study plans tend to include German, French, English and Italian. Of these, the first three are more intrinsically connected to art song (the requirement is to bring an art song in one or more of these languages for auditions, or to take a class in German, French and/or English song repertoire), whereas with Italian this connection is not so strong: I have not been able to find an ‘Italian art song class’ or ‘Italian art song’ audition requirement for any of these institutions; the requirement to sing in Italian is connected more to the vastness of the operatic repertoire in this language. A minority of the institutions also had classes or staff in Spanish, Russian and Czech. The eight concerts of art song at Wigmore Hall surveyed in footnote 1 included 67 songs in German, 40 songs in French, 20 songs in English and 1 in Russian. Oxford International Song Festival’s Texts & Translations portal (a leading resource for singers featuring the texts and English translations of more than 4,000 art songs, https://oxfordsong.org/discover) includes: 2410 songs in German, 1100 in English, 620 in French, 201 in Russian, 119 in Italian, 92 in Czech, 81 in Spanish. I therefore regard German, English and French as being the core languages in the linguistic regime of art song, with Russian, Czech, Spanish and Italian having secondary core status. I regard all other languages as lesser represented languages.
6
Werktreue is the expectation (typical of the Western art music canon) that a piece of music will be performed exactly as laid out in the score—which in turn is regarded as embodying the intentions of the composer.
7
Korean Art Song Resource, https://koreanartsongresource.com/
8
Ukrainian Art Song Project, http://ukrainiansong.art
9
Czech Art Song Initiative, http://www.czechartsong.com
10
The language needed to be unfamiliar to the singer, rather than ‘lesser represented’ in the art song context. While all singers reported having experience of singing in Italian, German, English and French, the choice was given to them to sing in, say, Russian or Spanish (to name two languages that are not particularly underrepresented in art song but might not be familiar to all singers).
11
For example the Song Database maintained by the Oxford International Song Festival, discussed above.
12
As a group, participants had received training in six different countries in three continents, and across a variety of settings, including conservatoire, university and private study. While their answers suggest that they were in general subjected to the same linguistic regime discussed in this article (for example, they all had experience of singing mostly in English, French, German, plus Italian for opera; they had all worked with language coaches), future research can help illuminate differences between countries or types of institutions. There were also different degrees of proximity between native and spoken language: Indo-European to non-Indo-European (two singers), and vice versa (one singer), plus two singers staying within the Indo-European language family.
Cite
DOI: 10.5871/jba/014.a19
Published on: 16 June 2026
Volume: 14
Issue: Issue 2
Article view count: 8
Article download count: 0
Copyright statement
© The author(s) 2026. This is an open access article licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International LicenseCite this article
Moreda Rodríguez, E. (2026), ‘Engaging with art song in lesser represented languages: expanding or challenging the linguistic regime?’, Journal of the British Academy, 14(2): a19 https://doi.org/10.5871/jba/014.a19Crossref
Dimensions
Altmetric
Article metrics
View count: 8
Download count: 0
No Data Found
No Data Found
No Data Found
Received
October 2022
Accepted
May 2023
Published
July 2023
Youth futures under construction Anna Barford
The articles presented here engage with some of the multifaceted and intersecting challenges faced by young people today – these include conflict, insecurity, limited government support, deep-set gender discrimination, climate change, infectious disease and a widespread lack of decent jobs. While recognising the structural influences on young people’s circumstances, the articles gathered here bring young people’s perspectives, experiences and actions to the fore. With an eye on the future, and a sense of the past, this collection is situated in the present. Most of the research presented here stems from the British Academy’s Youth Futures research funding scheme. The results showcased here remind us how the present matters in and of itself, while influenced by the past and playing a key role in shaping the future. Thus there is a triple significance to understanding young people’s challenges: they matter for today and for how they impact tomorrow, and will be best understood with reference to the past.
2 November 2023
Young people ‘making it work’ in a changing climate Anna Barford, Paul Magimbi, Anthony Mugeere, Mollen Nyiraneza, Benard Isiko, Charles Mankhwazi
Globally, young people face weak labour market demand and have been particularly susceptible to recent livelihood stresses and shocks linked to climate change. In this article, we consider what happens when young people face intersecting challenges including climate change. While much of the literature focuses on barriers to work and how to break these, we consider young people’s struggles and successes in securing and maintaining work. The focus is on Uganda, demographically one of the world’s youngest countries and home to a largely ‘underemployed’ cohort of young people. Our findings identify some of the many ways in which climate change disrupts young people’s livelihoods. Young people are already proactively responding to climate change. This points to the need for other actors to learn from young people’s existing endeavours, to build in more support and opportunities, manage risk and insecurity, and construct a more climate change-resilient infrastructure.
2 November 2023
No Data Found