An Example of Linguistic Stylization in Spanish Musical Genres: Flamenco and Latin Music in Rosalía’s Discography

Abstract

Studies on stylistic change in music argue that when singers use stylistic devices in their songs, they have one purpose: to represent their image and their artistic and social identity. In this paper we focus on the singer Rosalía, a Catalan artist who sings flamenco and Latin music. These two musical genres are associated with innovative varieties of Spanish, but this singer is not a speaker of this variety (she uses the Spanish spoken in Catalonia, a conservative variety). So, we want to know whether, when she sings, she linguistically adapts to the associated phenomena of flamenco and Latin music. In order to carry out this work, we have collected two oral corpora: the first is Rosalía’s discography and the second is 40 min of interviews in Spain and America. In our analyses we have verified that, indeed, when Rosalía sings, she uses indexicalized phenomena of the genres. However, in the interviews she keeps her vernacular variety, Catalan, although we have observed signs of accommodation to American Spanish, which also reveal a significant change in the singer’s idiolect.

1. Introduction

In this paper we provide a socio-stylistic and linguistic characterization of the singer Rosalía, a Spanish artist who currently has a great international projection. From a musical point of view, this artist performs in two different genres: flamenco and Latin music. Linguistically, these two types of music are related to two geolinguistic areas that use innovative variants of Spanish: flamenco, within Andalusia, and Latin music, within Latin America.

However, Rosalía was born in Catalonia (Spain); therefore, she does not have an innovative linguistic variety (this variety is mainly used in the southern part of Spain and in the Canary Islands). Rosalía is a bilingual speaker of Catalan and Spanish. When she speaks Spanish, she uses phonetic phenomena specific to the Spanish of the central northern peninsular and, in addition, linguistic characteristics from Catalan speakers of Spanish such as the palatalization of /-l/ at the end of a word, the voicing of /-d/ at the end of a word, which is pronounced as [-t] (verdad “true” as [beɾˈð̞at]) or the sonorization of /s/ in some speakers, who pronounce this sound as [z], e.g., in los amigos “friends” [lozaˈmiɣ̞os] (Garachana 2018; Villena-Ponsoda 2022; Poch Oliver 2019; Rius-Escudé 2020; Blas Arroyo 1995, 2004, 2021; García Mouton 2007; Vila-Pujol 2007). Therefore, the phenomena described above in flamenco are not typical of the vernacular variety of the singer.

In recent research on the linguistic characterization of singers in musical genres, the results have shown that professionals use certain indexicalized phenomena in each musical style, that is, linguistic features associated with particular music styles. Thus, singers change their accent to accommodate the requirements of the type of music and transform their vernacular variety and idiolect in musical contexts. Based on these general findings, in this research we pose a fundamental question: when Rosalía sings or speaks in music-related contexts, does she change her way of speaking? Does she use the variants that are recorded as representative of those musical genres? In this work, our main objective is to check the possible patterns of linguistic accommodation of the singer Rosalía to different communicative contexts: musical contexts, according to the genre she sings—flamenco or Latin music—and conversational contexts, especially in television interviews. Likewise, and based on this general objective, we want to observe the linguistic strategies used by the singer in the musical genres and also to analyze the phenomena she uses. On the other hand, and given that she has been living in America for some years now, it is possible that linguistic transfers between the two Spanish varieties (Catalan and American) are made; for this reason, we want to check whether these changes are indeed taking place beyond the linguistic adaptation to the musical genres and to determine whether there is a modification of Rosalía’s idiolect.

2. Musical Genres: The Object of Study of Socio-Stylistics

The study of speakers’ linguistic behavior is a socially and culturally relevant topic, because it allows us to understand linguistic variation from a group and individual point of view. When a person speaks, we can learn about many aspects related to their linguistic identity, their perception of their variety or, for example, their sociolinguistic awareness. However, when we listen to public personalities such as singers, we only know one version of them, namely, the version they want us to perceive. Following Bell and Gibson’s (2011) theories, singers are actors who can change their way of speaking according to extra-linguistic and socio-stylistic factors such as the audience they address, the musical genre they sing or, for example, the communicative context in which they speak. For this reason, the study of the linguistic behavior of singers is interesting: they have several linguistic profiles and choose which ones they use at different professional and personal moments.1

Research on the linguistic peculiarities of singers in musical genres is not new. In the 1980s, Trudgill (1980, 1983) did groundbreaking work on the linguistic characteristics of punk-rock; at this time, American English was the most widely used variety in pop music, but Trudgill noticed an important change: British bands used a hybrid accent with a mix between American and British English.2 A decade later, building on Trudgill’s findings, Simpson (1999) revisited the linguistic behavior of pop-rock singers of the 1980s and 1990s, and noted that this trend had indeed produced a liberation from American culture, and that songs used a mixture of the two varieties because of globalization. In fact, Simpson presented a model of English music (USA-5 MODEL, see in Simpson 1999, p. 345) in which he observed an effective register of phonetic phenomena that British singers use when singing in English. They retain their British variety but also use American phonetic variants. For example, Simpson records a change in vowels: he notes a shortening of the vowel [α] in words such as ask or last, which are pronounced as [a]; the change from [aɪ] to [a] in words such as life, my. Also, singers retain the prevocalic /ɾ/ as opposed to the British English vocalization and, finally, groups very often use the rhotic to pronounce the intervocalic /t/, as in water, which they pronounce [ˈwɑɾɚ] and not [ˈwɑtɚ].

Beal (2009) analyzed the linguistic behavior of the singer of the British band Arctic Monkeys, Alex Turner. She used the so-called USA-5 MODEL in performances, but she also used traditional and strongly localized phonetic variants. Actually, it was a strategy: the band refused to change their way of speaking, because they considered that these changes revealed a linguistic disloyalty to their variety. So, they decided to use traditional phonetic variants of their geographical area, Sheffield, which, moreover, related to young speakers. With this choice they represented their linguistic identity and their promotion of authenticity through the use of their own variety. It is true, however, that in some musical genres American English is very established; in fact, in most cases it functions as a predefined style of singing. For example, in the same direction, Gibson and Bell (2012) observed this pattern in New Zealand singers; they used American variants in their songs and concerts, but, in interviews, they significantly changed their pronunciation, while retaining their vernacular variety.

Other researchers have analyzed the linguistic characteristics of culturally charged music genres such as hip-hop and rap (Armstrong 2004; Edwards and Ash 2004; Samy Alim 2002; Gabarta 2023). In these studies, they have seen that American singers frequently use their variety (AAVE), although there are also some phonetic, morphological and lexical phenomena that are aimed at a particular audience, with variants that make it possible to represent a specific social reality. In these cases, although in music there is agreement on the linguistic phenomena of each musical genre (we will see this later in the section on the stylization of musical genres), singers are free to express their identity and their culture, and they are also free to address a specific audience. In fact, in current pop studies, Jansen and Westphal (2017) found that Rihanna sings in English and also uses Caribbean variants in her songs to represent her cultural origins; additionally, Eberhardt and Vdoviak-Markow (2020) have analyzed how Beyoncé represents her African American identity through morphosyntactic and pragmatic features in her songs, particularly since she released her album Lemonade in 2016.

In Spanish, from a general point of view, research on musical genres is still needed to observe and describe the linguistic behavior of Spanish speakers. However, in recent years there has been a growing interest in the study of the stylistic features of flamenco. For example, Fernández de Molina Ortés (2020, 2022a, 2022b) has found that flamenco, which is a genre that has an important place in the history of Spanish music culture, has associated phonetic phenomena. These sounds are related to the southern variety, used by speakers in the center-south of the peninsula and in the Canary Islands (see Fernández de Molina Ortés and Hernández-Campoy 2018) and, for more than a century, singers, regardless of their origin (southern or western) have preserved the stylistic phenomena of traditional flamenco.

2.1. Stylization and Indexicality in Music Genres

As we have explained in the previous section, a common tendency exists across musical genres: in pop-rock, hip-hop and also in flamenco some linguistic phenomena are related to each genre and, undoubtedly, singers use these features very systematically. Based on the studies we have previously discussed and also on their experience with the work on the Arctic Monkeys, Bell and Gibson (2011, p. 560) established four paradigms that explain how bands behave linguistically and also justify why each musical genre retains certain delimited characteristics: indexicality, stylization, register and discursive culture.3

As we have seen, in musical genres there are linguistic features (phonetic, morphological, lexical or pragmatic) that are associated with each style. We saw this, for example, in British pop-rock: some hybrid phonetic phenomena between British and American English (USA5-MODEL) became sounds of this type of music and the singers, over time, have preserved them. So, since there are indexical features of a style of music, there is a stylization of that genre: the phenomena represent the music, and the singers linguistically accommodate those features, because they are considered representative of that way of singing. In fact, this is especially important. As Bell and Gibson explain, the public registers these linguistic phenomena in its mind; we recognize a rock singer by a particular aspect, and we recognize an English rock singer because they employ some phonetic phenomena. These registers are in the public mind and represent styles of music or cultural and social styles (Agha 2003). However, while it is true that the indexicality of linguistic phenomena in musical genres is evident, music styles change over time and regenerate; this was found by Trudgill (1980, 1983) in early studies and is also being observed by researchers in British pop and rock today (Jansen 2022).

In Spanish, as we indicated above, these four paradigms are represented in flamenco. In fact, according to the latest research (Fernández de Molina Ortés 2020, 2022b), when flamenco became popular in the late nineteenth century, a phonetic stylization of the genre occurred.

It is necessary to explain that this musical genre appeared as a private musical activity performed by Andalusian gypsies in their homes, with their families and friends. But, from 1880 onwards, these performers, the cantaores, became professional, and began to sing in cafés cantantes (“coffee with music”) in Andalusian cities. These shows were successful, and the cantaores traveled to other parts of Spain to sing flamenco. The way of singing of these new professionals was a model for representing flamenco; these gypsies, who were Andalusians from the western area, transferred the phonetic phenomena of their area with their singing. From 1920 onwards, flamenco became very successful in Spain; during those decades, new non-Andalusian cantaores (Madrilenians, Catalans, Castilians) were formed, who learned from the Andalusian cantaores. They imitated this way of singing, and progressively the linguistic features became indexicalized, that is, the features became recognized as representative of the flamenco style.

Nowadays, any fan of flamenco singing knows the phenomena of singing: the seseo, the fricatization of [tʃ] which is pronounced as [ʃ] for the spelling ch ([muˈʃaʃo] muchacho “young man”), the lenition of articulatory sounds. For example, the implosive /s/ is pronounced aspirated ([me ˈɣ̞uhta] me gusta “I like it”) or dropped ([loh ˈβ̞eso] los besos “the kisses”); the voiceless velar fricative/x/is also aspirated in the initial position ([ˈdʒo ˈsoi̯ hiˈtano] yo soy gitano “I am gypsy”) and interior ([ˈkohelo] cógelo “take it”). We can also include rhotacism as representative of gender, i.e., the use of [ɾ] and not [l] when speakers pronounce el niño “the child” as [eɾ ˈnῖɲo] or calma “calms” as [ˈkaɾma].

However, as in other musical styles, flamenco changed; in the 1960s, new musical genres such as jazz, blues and rock arrived in Spain. Cantaores mixed these new rhythms with flamenco and created “flamenco fusion”. This was a natural evolution of the cante because, as Steingress (2005, p. 133) explains, there was a “logical continuation of the transgressive and transcultural character of flamenco itself” which, although with a traditional character, “was transformed into a modern and urban manifestation of popular art”. In this variety of flamenco, we can observe some phonetic changes, especially in non-general variants in southern varieties such as the fricatization of /tʃ/ or rhotacism, for example. However, the essence of traditional flamenco is still preserved in this variety of cante (Fernández de Molina Ortés 2022a).

3. Materials and Methods

3.1. Who Is Rosalía?

Rosalía Villa Tobella is a singer of Spanish origin, born in San Cugat del Vallés, Barcelona, in 1992. She is an artist who has developed her artistic career in two parts: at the beginning of her career, Rosalía became known with flamenco fusion albums. In 2017 she released Los Ángeles and, a year later, El mal querer (2018). This album received two Grammy Awards (the Latin Grammy for album of the year and the Grammy for best urban or Latin rock album). The second part of her career began in 2019. Rosalía started to collaborate with some Latin music singers: J. Balvin (“Con altura”), Ozuna (“Yo x ti tú x mí”) and Billie Eilish (“Lo vas a olvidar”) and, today, she is a singer of urban music. Since 2019, she has been living in the United States of America, where she recorded her latest album Motomami, which she released in 2022. In this album she mixes flamenco, urban and Latin bases.

3.2. Corpus Development and Sample Determination

In order to analyze the phonetic phenomena of the singer Rosalía, a complex oral corpus has been compiled and analyzed, divided into two categories: musical and public.

In the musical corpus, as we can see in

Table 1. Summary of the oral corpus configuration.

Type of Intervention		Example	Time
Musical intervention	Flamenco	Los Ángeles (2017)	49:06 min
		El mal querer (2018)	28:10 min
	Latino	Motomami (2022): “La fama”, “Candy”, “Delirio de grandeza”, “Bizcochito” Colaboraciones: “Con altura”, “El pañuelo”, “Yo x ti tú x mí”, “TKN”, “Relación Sech” “Lo vas a olvidar”, “La noche de noche”, “Besos mojados”	34:38 min
Interviews	National	Interview on Radio 3 (2022) I	10 min
		Interview with Javi y Mar (Cadena 100) (2022) II	10 min
	International	Interview with Alofoke (Dominican productor) III	10 min
		Interview “Primer impacto” (Univisión) IV	10 min

^I Interview is available here: https://www.youtube.com/watch?v=g-kx2bkzIxY&t=569s (accessed on 10 December 2022); ^II interview is available here: https://www.youtube.com/watch?v=xI5f5CtnAE8&t=44s (accessed on 12 December 2022); ^III interview is available here: https://www.youtube.com/watch?v=pihmCIbsgAI (accessed on 12 December 2022); ^IV interview is available here: https://www.youtube.com/watch?v=rCuu6w_7Xw0 (accessed on 9 January 2023).

, we have subdivided the singer’s discography into two different genres: flamenco and Latin music. With this classification we represent the two main stages of Rosalía’s discography, which has undergone a great musical transformation.

On the other hand, we have created two subcategories to analyze Rosalía’s interventions in the public sphere: first, we present the interviews she conducted in her country (Spain) and, second, the interviews in the international sphere, specifically in America. We have chosen four interviews (two for each geographical area), with a total analysis of 40 min. The interviews have a similar format: in the national context, we have selected two recordings where Rosalía promotes the album Motomami; in the international context, we have chosen two programs where the singer is interviewed personally.

Finally, we have to point out that we have compiled two control corpora in order to compare the linguistic characteristics of the singer with the musical genres we are going to analyze (flamenco and Latin music).

As can be seen in

Table 2. Features of the control corpus.

Style of Music	Analysis Time	Analysis Lemmas
Flamenco	44 h	94,978 lemmas
Latin music	47:22 min	1401 lemmas

, the flamenco control corpus is composed of 44 h of recordings. This has allowed us to compile a large dataset (94,978 lemmas) with different linguistic features. This corpus is composed of 684 traditional flamenco cantes and flamenco fusion cantes.4 Rosalía studied traditional flamenco in the course of her academic studies, and we assume that she uses the phonetic phenomena of this type of flamenco. However, her music is not considered traditional flamenco. In fact, it is more related to flamenco fusion. With the data of the control corpus, we can see which flamenco the singer has as a reference.

On the other hand, we have compiled the control corpus of Latin music, which is composed of 46 min of recordings of songs from the last decade. We employed a random selection process to choose songs from the corpus. We have analyzed fewer lemmas than in traditional flamenco (1401 words), but we have used a variety of singers from different areas to uncover the general linguistic phenomena of this musical genre.5 Latin music has not yet been exhaustively analyzed, and therefore this small musical base will allow us to study its main features; later we will be able to compare it with Rosalía’s interpretations.

3.3. Linguistic and Extra-Linguistic Variables

As indicated in the introductory section of this work, we want to investigate whether Rosalía adapts, that is, if she accommodates her language to the phonetic characteristics associated with certain musical genres and the linguistic diversity of her interviewers. To begin with, we can affirm that Rosalía belongs to the linguistic variety typical of the centre-north of the peninsula, because she was born in Catalonia. According to this geolinguistic origin, the singer maintains the articulatory characteristics of the sounds in the implosive and final position (/-s/, /-d/, /-θ/), maintains the friction in the voiceless velar /x/ and pronounces the intervocalic /d/. In addition, as is typical of the Spanish spoken in Catalonia, we can find the pronuntiation of /-d/ as a voiceless sound /-t/, the sonorization of /-s-/ as [-s-] or the palatalization of /-l/ in the -al ending. Likewise, it could be said that she does not use Spanish phenomena such as seseo (typical of some areas of Andalusia and the American variety) or the fricatization of /tʃ/, for example, associated with Lower Andalusia. Finally, in her geolinguistic area, it is also unusual to find such phenomena as rhotacism or lambdacism, the aspiration of the h coming from the Latin F-.

However, as we saw in the theoretical section of this work, although singers do not have some phenomena in their variety, they know them and can use them as a representation of the musical genre. Flamenco and Latin music are genres with a long tradition and some phenomena are indexicalized. Thus, for example, to sing flamenco, singers use variants of southern Spanish and also some phenomena such as seseo, the fricatization of /tʃ/ and even rhotacism. In Latin music, professionals use similar features (we include here, too, seseo) and, in addition, some singers produce a change in liquids with a result of lambdacism6 due to the influence of the Caribbean variety.7

In this work, the linguistic variables we use to check Rosalía’s linguistic behavior are phonetic:8

• Articulatory lenition of the /s/ sound. In Spanish, the pronunciation of the sound /s/ in the implosive and final position presents some variants produced by articulatory lenition. The /s/ can be pronounced aspirated ([ehˈtɾeʝa] estrella (“star”)) or elided ([ˈfloɾe ˈasule] flores azules (“blue flowers”)).
• Aspiration of /x/ (velar fricative voiced sound). In some innovative Spanish-speaking areas /x/ is often pronounced as [h]. Thus, it is possible to hear an utterance like iré joven (“I Will go Young”) as [iˈɾe ˈhoβ̞en], for example.
• Seseo. Neutralization of the sounds /s/ and /θ/ with an [s] solution. In Rosalía, for example, we have found examples of seseo when she says [dehkonoˈsia] for desconocida (“unknown”).
• Neutralization of liquid sounds /l/ and /ɾ/. In this variable we are going to analyze two different results according to the type of musical genre. As Fernández de Molina Ortés (2020) found, in flamenco we can find examples of rhotacism, i.e., a neutralization of /l/ and /ɾ/ in favor of the rhotic [ɾ]. Thus, an example of rhotacism would be to pronounce alma (“soul”) as [ˈaɾma], for example. In Latin music, especially in Caribbean singers, we can find examples of lambdacism, i.e., pronouncing /ɾ/ as an [l]. For example, a case of lambdacism would be to pronounce comer (“eat”) as [koˈmel] (López 2016; Moreno-Fernández 2019; Maymí and Ortiz-López 2022; Moreno Fernández and Otero Roth 2016).
• Elision of intervocalic /d/. In Spanish, the voiced stop /d/ within a word is pronounced as an approximant [ð̞]. In this intervocalic context, /d/ undergoes articulatory lenition and, in some cases, the sound disappears completely. In Spanish, it is common for the elision of /d/ in the ending -ado (cansado “tired” [kanˈsao]). However, in other endings it can be lost, as in -ido, -ida (perdido, perdida “lost” [peɾˈð̞io], [peɾˈð̞ia]), in -odo and -oda (todo, toda “all” [ˈto], [ˈtoa]). In flamenco, it is even found in some combinations as -eda (enfermedad “illness” [ẽɱfeɾmeˈa]). However, the loss of /d/ is a phenomenon related to medium and low sociolects.
• Other changes. From a phonetic point of view, we analyze syllabic and vowel apocope in words such as para (“for”) and muy (“very”). In Spanish these are pronounced as [ˈpa] or [ˈmu] in relaxed contexts; we will also analyze the loss of intervocalic /d/ in different endings.

In this work we have also checked the possible influence of some morphosyntactic variables typical of American Spanish in the singer, and we will especially focus on subject–verb constructions (esa yo soy “that’s me”). In addition, we observed that Rosalía uses the intensifier bien where, in European Spanish, muy (“very”) is generally used.

Finally, we believe that it is also interesting to know the use of lexical variants that the singer has incorporated into her way of speaking; in this variable we will include two types of variants: Americanisms, that is, words whose meaning is associated with an American reality; and Anglicisms, that is, words of English origin that the artist uses in her conversations in Spanish.

On the other hand, and as shown in Section 3.1, this research also includes some extra-linguistic variables, mainly stylistic. Thus, we have analyzed whether the type of musical genre (flamenco or Latin music) influences the Catalan artist’s linguistic accommodation to the indexical features of each genre. Likewise, we also observe whether the place of the interview (national and international) favors a greater linguistic adaptation of Rosalía to her interlocutors.

4. Results

Before starting a detailed analysis of the data, we would like to show the linguistic behavior of Rosalía in the corpus we have analyzed. In total, 968 variants have been used for this general analysis and two categories have been established: non-innovative variants and innovative variants. We used three different phenomena for grouping: the pronunciation of the implosive /s/, the pronunciation of /x/ and the presence or absence of seseo. As the singer is Catalan, the aspirated or elided pronunciations of /s/ and /x/ and the use of seseo are the innovative variants because, as indicated above, they do not belong to her linguistic variety.

According to the general results of our corpus(

Table 3. General features of the pronunciation of sounds (innovative and non-innovative) by Rosalía.

		Musical	Public	Total
Non-innovative	N	52	393	445
	/%/	9.7%	97.5%	100.00%
Innovative	N	513	10	523
	/%/	90.8%	2.5%	100.00%
Total		565	403	968

), which have statistical significance (χ²: 738,647 (1) < 0.001), we could affirm that there is a pattern of linguistic accommodation in the musical context, where in 90.8% of the cases Rosalía uses innovative variants that do not correspond to her variety. However, in the interviews, she does retain phenomena typical of the Catalan variety (97.5%). Therefore, according to these data, we begin the analysis of the corpus knowing that there are two very different tendencies of accommodation.

4.1. Linguistic Uses in the Musical Sphere

As we saw earlier (Section 4), Rosalía uses innovative variants in the musical field; in fact, according to our data, the frequency of use of innovative phonetic phenomena in this context is 79.10% in flamenco and 79.50% in Latin music. These results show that the use of innovative phenomena is similar in both genres, although it is necessary to check which phenomena are used in each type of music in order to establish the most representative characteristics of the singer when she sings in each style.

4.1.1. Flamenco

As shown in Section 2, the linguistic configuration of flamenco is represented by its own phonetic phenomena. Since the beginning of the genre, some sounds have been associated with cante and, although most of them are related to phonetic phenomena of the southern areas, others are related to variants of Andalusian speech (such as fricatization and seseo). Other phenomena such as rhotacism or the drop of the intervocalic /d/ are also included.

If we check the phonetic phenomena of traditional flamenco and flamenco fusion in our control corpus and Rosalía’s use of these variants, we can affirm that there is an evident accommodation of the Catalan singer to the phenomena recorded in traditional flamenco singing. On the one hand, and as we observe in Figure 1, Rosalía frequently uses the articulatory lenition of the implosive and final /s/ (94.9%), the aspirated variant of /x/ (89.7%) and the seseo (79.2%). The first two phenomena are widely used in the innovative varieties and have been features of flamenco since its origin. Seseo, on the other hand, is a more geolinguistically localized phenomenon and is associated with the western part of Andalusia.9 This variant has also been used as an indexical feature of flamenco singing from the beginning and, moreover, enjoys linguistic prestige in current Spanish (see Carbonero Cano 2007).

On the other hand, we will analyze the group of other features presented in Figure 1: the fricatization of /tʃ/, the elision of intervocalic /d/ and rhotacism. These variants have been recorded in the flamenco control corpus because, as Fernández de Molina Ortés (2020, 2022b) observes, these features were indexicalized in the origins of flamenco singing and have been recorded in all periods (from 1880 to the present day, see Section 2.1). However, they are different uses from the previous phenomena because, for example, some of them, such as the loss of /-d-/ or rhotacism, are not exclusive to the southern variety. Likewise, the fricatization of /tʃ/ is a variant associated with the western area of Andalusia, but it is limited to some geographical areas; moreover, in urban areas it is related to medium and low sociolects.

However, if we observe how Rosalía uses these phenomena, we can see that she chooses all the variants very frequently, even more so than the singers of traditional flamenco. The artist pronounces the fricative in more than half of the data (55%/62.5%), and she used rhotacism with great systematicity (63.5%/78.9%). The elision of the intervocalic /d/ is also striking (97.5%) because Rosalía uses this variant constantly and in any ending (-ada, -oda, -ido, -ida, -eda, etc.).

These results are interesting: from these data we can say that Rosalía frequently uses southern variants in her songs to identify herself with flamenco; but, above all, what is most relevant is that the singer reinforces the indexical phenomena of traditional flamenco that have more representative (geolinguistic and social) marks, with the aim of underlining the essence of this genre in her music. The flamenco albums we have analyzed, Los Ángeles and El mal querer, are ascribed to flamenco fusion, but the singer chooses phonetic variants of traditional flamenco. In this way, we can confirm that Rosalía knows the singing phenomena of all periods and systematically chooses variants associated with tradition so that, linguistically, the fans associate her way of singing with this type of singing. We have also found other interesting phenomena in Rosalía’s songs that complement these conclusions. For example, we have observed examples of syllabic apocope in the preposition para “for”, “to”, which Rosalía pronounces as [ˈpa] (18 cases). She also uses this phenomenon in the adverb muy “very” as [ˈmu] (17 cases), for example, in the chorus of “muy mal, muy mal, muy mal, muy mal, muy mal” (“very bad”) in the song “Augurio”.10

Finally, it should be noted that Rosalía uses a variant typical of the new generations of flamenco: the labiodental /v/ for the labial /b/. We have compiled 17 cases of pronunciation of this sound in examples such as besos (“kiss”) [ˈveso], olvidar (“forget”) [olviˈð̞a] or abanico (“fan”) [avaˈniko].11 We have also noticed examples of articulatory reinforcement of the group [st] with an affricate pronunciation [hts], which is increasingly frequent, especially in western Andalusia (Ruch and Peters 2016; Ruch 2008; Vida Castro 2015, 2016, 2018, 2022; Moya Corral 2007). In studies of traditional flamenco and flamenco fusion this variant also appears in younger generations, and Rosalía continues this trend in her songs.

4.1.2. Latin Music

In order to check whether, as in flamenco, Rosalía adapts to Latin music, we have compiled an oral corpus and we have analyzed three phenomena that have an important variation in innovative Spanish: the pronunciation of the implosive /-s/, the voiceless velar fricative /x/ and the variants of the intervocalic /d/; in addition, we have analyzed the pronunciation of /θs/ (and, in this case, the presence of seseo) and the neutralization of liquid /l/-/ɾ/, specifically with lambdacism.

Figure 2 reveals that, indeed, according to our control corpus, phonetic variants typical of innovative areas of Spanish are frequent in Latin music: aspiration and elision of /s/ (75.1%), aspiration of /x/ (96.6%) and the pronunciation of /θ/ as [s], i.e., the use of seseo (100%). Likewise, the singers choose the elision of the intervocalic /d/ (59.4%), although it is not a systematic use (we find here, therefore, a difference with flamenco).

On the other hand, from the control corpus, we have observed that some singers use lambdacism, and we have found that this phenomenon has a frequency of 60.70% (it is used in 42 out of 112 occurrences).

Initially, we could include this linguistic feature as a variant in Latin music. However, with a more detailed analysis of the data, we have found that lambdacism is not an indexical phenomenon of this genre. As can be seen in

Table 4. Use of lambdacism in Latin music singers of Spanish-American origin.

	Caribbean		Non-Caribbean
	N	%	N	%
Lambdacism	41	63.1%	1	2.1%
Non-lambdacism	24	36.9%	44	93.6%
Assimilation	0	0%	2	4.3%

, it is a phonetic feature used by singers who come from linguistic areas with lambdacism, specifically, from the Caribbean area (63.1%); however, singers from other areas do not choose this variant, but pronounce the rhotic /ɾ/ as [ɾ].

From a general description of the phenomena associated with Latin music, we can check Rosalía’s linguistic behavior in this genre (see Figure 2). First, we can affirm that Rosalía uses in a similar way the phenomena that she had already used in flamenco singing in her initial stages as a singer; in fact, the frequency of the lenition of /s/ (84.3%) and the aspiration of /x/ (95.5%) is very similar to the singer’s flamenco corpus. Rosalía also retains seseo, a phenomenon recorded in cante and Latin music, given that most singers have an American origin, where seseo is a systematic variant. However, if we look at the data on the elision of intervocalic /d/, we do find significant changes: in Latin music, Rosalía uses elision with great systematicity (91.7%); this percentage differs from the results for /d/ in Latin music (59.4%) and is more similar to the data we observed in flamenco (see Figure 1).

On the other hand, in this musical style, Rosalía does not produce any neutralization of liquid sounds, that is, she does not use phenomena such as rhotacism and lambdacism; therefore, she keeps the opposition of /l/-/ɾ/.12 This differs significantly from the phonetic changes that the singer made in flamenco, where she did use rhotacism with a high frequency (78.9%). At first it could be argued that, like other singers of non-Caribbean origin, she does not consider it relevant to use lambdacism in her songs because it is not an indexical phenomenon of Latin music; in flamenco, on the other hand, rhotacism is considered a stylistic feature of the genre; in fact, she uses it more frequently than other members of the traditional flamenco corpus.

However, it should be noted that, although Rosalía does not choose lambdacism in Latin music, 15 occurrences with some representative changes in the pronunciation of /ɾ/ and /l/ have been noted in the corpus. Thus, for example, we have observed an assimilation of /ɾ/ to the following sound. This is heard, for example, in [ˈdʒo poɾ ˈti ǀ ˈtu poˈmi], in the song “Yo x ti, tú x mí” and in the chorus of the song “La noche de noche”, where she says [poke la ˈnotʃe ǀ ˈnotʃe ǀ ˈfwe]. These pronunciations may represent a phonetic change with which the singer is accommodating not only to the phenomena of Latin music but, specifically, the collaborators with whom she performs these songs. In both cases, the singers come from Puerto Rico (Ozuna and Bad Bunny) and, when they sing, they generally use lambdacism. In fact, this change corresponds to a pronunciation typical of Caribbean areas.

In addition, Rosalía uses other phenomena that are associated with Latin music and also with flamenco. We are referring, for example, to the syllabic apocope in the word “para” (she uses this variant in 16 cases). This feature can also be considered general in the Latin style, because in the control corpus we observed that in 90% of the examples collected, the singers also use the apocopated form.13 We have also found some examples of syllabic apheresis in the pronunciation of [ˈtamos] and not [es̪ˈtamos] (“we are”), a resource which, in fact, also appears in some singers such as Becky G. or Karol G., for example.

Finally, and from the lexical point of view, we observe that Rosalía’s songs combine European Spanish with American Spanish through the use of Americanisms such as “estar a machete”, “lo tengo juqueao como si fuera tote”,14 “el dembow”15 or the use of “el cora” as a variant of “corazón”.

Likewise, and more frequently, the singer uses Anglicisms in her songs (21 examples have been collected). This resource seems to be frequent in Latin music, because in the control corpus we also collected 118 occurrences. For example, Rosalía uses the variant “baby” as a vocative on four occasions; this expression is also very frequent in other Latin music songs in the control corpus. However, it is more interesting to observe the combinations she makes in some verses: “estamos Worldwide a machete”, where we notice the presence of Americanisms and Anglicisms. It is also interesting to note the combination of the two languages in the songs, as well as the pronunciation of English words in Spanish. This can be seen in some verses, such as: “ni un amigo nuevo, ni uno haría zoom en la cara” or “millonaria con hoodie barata”, “todo me queda bien con una White tee” or, in the following verse, “todos me quieren en casa, soy una wifey”, “lo subo al cielo, yo soy su messiah”.

4.2. Linguistic Characteristics of the Interviews

In previous sections we have observed that Rosalía does indeed adapt to the indexical phenomena of musical genres. However, does she use the same strategies when conducting interviews in different geolinguistic contexts? In Table 2 we observe that there was a smaller percentage of accommodation; in the following, we will check in which areas she chooses innovative phonetic variants.16

Based on the data in Table 4 from the analysis of innovative phonetic variants, we can affirm that we do not find linguistic accommodation in the interviews she conducts for the Spanish and American press. In fact, only 8.1% of innovative results were found in the international sphere. In most cases, she does not make systematic changes; for example, we have observed that Rosalía uses the seseo on one occasion (when she says [sjeˈn poɾ ˈsjen] “cien por cien”) and, on five occasions, she uses the lenition of /s/. However, we have found a relevant change in Rosalía: as in Latin music, in the American interviews there is also an evident change in the pronunciation of the liquid sounds /l/ and /ɾ/. In her answers to the interviewers, Rosalía uses heterogeneous solutions: on two occasions she chooses an assimilated /ɾ/, when she pronounces [poˈnelo] (ponerlo “put it”), [ˈpake] (parque “park”); we have also found three lambdacisms in the word “hermana”, which she pronounces [elˈmãna] (hermana “sister”). This usage is very representative, because it is a phenomenon that does not appear in music but does occur in face-to-face communication in these contexts. And more interesting is the pronunciation of the word enamorarme as [ẽnãmoˈɾaɻme] (“falling in love”), where Rosalía pronounces a rhotic as a retroflex sound /ɻ/, a variant which is not present in European Spanish. It is an American English sound that is often used in Puerto Rico. These changes are not systematic (they are very infrequent), but we believe that they do reveal a possible accommodation of Rosalía to a new linguistic space, the Spanish-American one. We also believe that the changes in the liquid sounds, which reflect different phonetic solutions, may also be related to the direct contact she has with her current partner, Raw Alejandro. This singer is of Caribbean origin (from Puerto Rico), and, in fact, it is possible that the lambdacism in the word hermana [elˈmãna] is related to the frequent use of this variant in the family environment.

On the other hand, and from a morphosyntactic point of view, some interesting accommodations of the singer have also been observed during her American interviews. The most representative use is the combination of the adjective “bien” as an intensifier with an adjective. Although in European Spanish we can find “bien” in some combinations such as in “dejarlo bien claro” (“make one thing very clear”), “llegó bien tarde”17 (he/she arrived very late) with the adjective bien as an intensifier together with an adjective, the use Rosalía makes is undoubtedly related to American Spanish:

• Me siento bien agradecida;
• Yo pienso en que casi todo lo que me va pasando es bien orgánico;
• El pueblo tiene un nombre bien, bien largo.

In fact, if we observe the frequency of use of this intensifier in these contexts in the databases of the Real Academia Española (CORDE and CREA), most of them come from American areas; example 1, “bien agradecida”, is present in Venezuela and example 3, for example, is also registered in Venezuela, Chile and Cuba.18 Likewise, a transposition of elements has been observed in the syntagm “yo soy de ahí”; to explain its origin, Rosalía says: “al este de ahí es que yo soy”, where the subject is placed before “yo soy”, a construction typical of the Caribbean variety (see López 2016; Moreno-Fernández 2019). Finally, at the lexical level, some examples have also been observed:

4.

Me gusta la gente, la vibra aquí, en la calle;

5.

¿Crees en el matrimonio? Sí, definitivo;

6.

¿Le gusta el snapple?;

7.

Me acuerdo de cantar en los shows.

In these examples we observe Americanisms such as the word “vibra” (4) (a diminutive of “vibration”) or the expression “definitivo” (5), which is used as a synonym for “absolutely”. Other words are related to brand names (snapple) (6) and, also, on seven occasions, she uses the term “show” to talk about her performances (7). These uses are different from the songs: Rosalía uses some American terms but not frequently.

5. Discussion and Conclusions

In the results of this investigation, it has been demonstrated that, effectively, in musical genres, singers use some strategies to represent the characteristics of the musical style and, also, to represent their own linguistic identity. Rosalía is an example of this linguistic behavior: she uses stylistic phenomena of these genres and adapts linguistically when she sings. It is interesting to note that when she sings flamenco and Latin music, she chooses innovative variants that do not belong to her vernacular variety. With this, she creates in her idiolect a new variety, the musical one. Considering her biography, we consider that Rosalía first learned the flamenco variety and used innovative variants of Spanish which, until now, were not integrated in her way of speaking. In fact, it is interesting to note that this singer uses sounds from southern areas even more frequently than the Andalusian singers of traditional flamenco. In her interpretations she therefore reinforces linguistic features that are traditionally closely related to flamenco. Subsequently, the phenomena that she knows from flamenco, she transfers to Latin music. In this musical genre, the singers use innovative variants of American Spanish, which mostly correspond to those of European Spanish. It is easy for Rosalía to use these kinds of variations in the musical context because she has used them before in flamenco. From a stylistic point of view, she uses the stylistic phenomena of each genre, choosing indexicalized variants that are in the cultural register of fans.

In interviews, however, this singer frequently retains her Spanish-Catalan variety; we have only observed some changes in international interviews. Therefore, we could say that Rosalía does indeed code-switch when she sings but, in face-to-face conversations, she retains her variety.

However, although the changes occur mainly in the musical context, we have been able to see some signs of variation in the singer’s personal idiolect. In the international sphere (in music and in interviews) we observed that Rosalía uses different variants of the sound /ɾ/: assimilation, neutralization with lambdacism or even a retroflex pronunciation, typical of American English and also of Puerto Rican Spanish. These changes are a representation of the changes in the singer’s idiolect: she is in direct contact with the American variety, especially the Caribbean one, and her accommodation to it is evident in certain features. Evidently, Rosalía has a great facility for linguistic adaptation, as we have already seen in the musical genres. When she speaks in international contexts, she uses the same process of accommodation, although less frequently. Rosalía is an example of linguistic change in Spanish musical genres. She is aware of the paradigms of these communicative contexts and generally complies with them. As observed in research on musical genres, this behavior of singers is frequent as they accommodate to the audience and the perception that audience members have of a musical genre. In Spanish, we have still to see how the idiolect of singers has evolved, carrying out real-time studies that allow us to describe specifically the Spanish music of recent times.