Introduction & Background

Universities across the U.S. are linguistically diverse spaces, but on a campus as diverse as UCLA, students regularly navigate between multiple languages in their everyday interactions. English often functions as the dominant institutional language in academic and public settings, leading our group to pose the question: how do bilingual students decide when to switch between languages? The motivation for this study comes from the idea that language choice is not random. While previous research shows that code-switching expresses identity, our group seeks to highlight how bilingual students regulate their language use in response to social and institutional expectations. This study explores this gap, showing how these students adjust their language choices to navigate identity and structural power in everyday campus settings. We argue that multilingual speakers strategically regulate code-switching across private and public settings in response to audience comprehension and institutional language norms, revealing how language choice negotiates identity, accessibility, and structural power.

Methods

This study examines code-switching among bilingual undergraduates at UCLA across diverse social contexts. The study focuses on 14 undergraduate students at UCLA who speak two or more languages and report having used them for a long time. These languages included Armenian, Farsi, Spanish, Mandarin, Korean, and Arabic. Participants were selected through convenience sampling from the researchers’ personal networks, consisting of a close-knit group of friends. Thereby facilitating access to individuals familiar with the campus environment and daily student interactions, both in private and public settings. The research examines both public settings (such as classrooms, campus areas, and academic interactions at UCLA) and private or informal contexts (including conversations with family and close friends, within friend groups, and dorm rooms). This method aligns with the principles of linguistic anthropology, which examines daily interactions as a fundamental site for constructing social meaning, in which language use reflects broader cultural and social dynamics (Dingemanse & Floyd, 2014).

Short questionnaires and casual interviews are used to get participants to talk about their own language use. The analysis does not focus on formal linguistic structure; instead, it examines observable elements in communication, such as code-switching, participants’ identities, and motivations for switching, including the expression of identity, humor, emotion, or clarification. This method is based on conversation-based techniques that examine how to design an audience, how to take turns, and what people mean when they talk to each other in real life (Bahtina, 2026A).

Patterns are identified in all participants’ responses by categorizing analogous behaviors and interpretations. This is a common way to link participants’ social contexts to their interactions (Bahtina, 2026B). Close attention is paid to how language use changes across different settings, such as family, friends, and school, and how these changes reflect societal and institutional expectations for how people should act (Bahtina, 2026C).

The objective of this methodology is not to produce results that could be generalized at a statistical level; rather, it is to uncover significant patterns in the manner in which UCLA bilingual students govern language usage in everyday campus life, as well as the ways in which these choices reflect identity, social connections, and structural factors. Therefore, the goal of this methodology is to produce generalizable results.

Results/Analysis

Figure 1. What comes to mind when students talk about code-switching. This word cloud is based on responses from 14 UCLA bilingual students, with larger words reflecting ideas that came up most often.

The following quotes are drawn from our data and grouped by recurring themes in students’
code-switching experiences.

1. Identity
   “Switching between languages is just part of who I am.”
   “It reflects both sides of my identity.”
   “I feel like I’m a different version of myself depending on the language I use.”
2. Community & Belonging
   “When I hear other people switching, it makes me feel like it’s okay for me to do it too.”
   “It creates a shared understanding between us.”
   “It makes conversations feel more natural and comfortable.”
3. Humor & Expression
   “Some jokes only work in one language, so I switch.”
   “It’s the only way to make it sound right.”
   “Certain phrases just hit better in another language.”
4. Emotion
   “There are things I can say in one language that don’t feel the same in another.”
   “I feel more comfortable expressing emotions in one language.”
   “Some emotions are easier to express in my other language.”
5. Institutional Pressure
   “In class, I usually stick to English—it just feels expected.”
   “It feels more appropriate to use English in academic settings.”
   “I don’t want people to think I’m unprofessional.”
6. Naturalness
   “It just happens naturally—I don’t really think about it.”
   “Switching languages feels automatic.”
   “It’s just how I talk.”
7. Audience Awareness
   “It depends on who I’m talking to.”
   “I switch based on whether the other person understands.”
   “I adjust my language depending on the situation.”
8. Family vs Campus
   “At home, we usually stick to one language.”
   “On campus, I switch more because people are bilingual.”

Code-switching socially during in-person communication is frequent in conversations with bilingual friends, casual conversations, joking or storytelling, and emotional conversations. This was connected to people hearing others code-switching, which encourages them to do so as well; it creates a shared understanding, feels more natural, and signals that both languages are acceptable, per participants’ answers.

More code-switching occurs on campus because students come from diverse linguistic backgrounds and many are bilingual. Shared spaces including dorms, restrooms, and study rooms help students be comfortable switching between languages. As one participant explained, “Some academic terms feel more natural in English, while certain emotions or cultural references feel more authentic in Chinese,” providing more information behind when and why students code-switch. At home, families often speak one main language, so there is less switching and more focus on that language. This contrast suggests that code-switching is shaped by one’s linguistic environment and audience (Zentella, 1997; Duranti, 1997).

English is used primarily in academic and professional settings, while other languages are mainly used in family and cultural environments (Gumperz, 1982; Zentella, 1997). Code-switching, however, occurs more often in casual social situations, especially with bilingual friends who speak the same language. Hearing others code-switch often makes people more comfortable doing it themselves. A participant in our study wrote, “At work and during clinical, I also code switch when helping translate for staff who don’t speak Spanish and patients who do. Being able to move between languages is important because it helps with communication and makes patients feel more understood and supported. It’s not just social, it’s something that has a
real impact in healthcare settings.” This shows the power of language and the role that codeswitching plays in and out of social communities, crossing institutional barriers and structural power. It also conveys how code-switching functions not only as a social practice, but as a communicative resource, allowing speakers to build trust and navigate systems where language differences might otherwise create misunderstanding.

Discussion & Conclusion

In healthcare and legal settings, our findings suggest the possibility to prevent “linguistic profiling,” ensuring that patients and defendants are understood based on their meaning rather than judged by their accent or dialect. Our data also portrays when bilingual students utilize code-switching it provides a further sense of support and cultural connection, especially in academic spaces where students may feel underrepresented. Future studies can sample a larger population in different locations to further study language choice and narrow the study to specific locations such as, restaurants, office workplaces etc.

By recognizing that every word we use has power, we can move beyond stereotypes towards a deeper shared understanding. We show that our words are never just about grammar; rather, they are primary sites of identity construction and a form of cultural contestation. We move beyond simple definitions to reveal that alternating between languages and combining languages, using dialects and different registers, are not barriers or signs of deficiency; they are high-level communicative competencies used to navigate the social world. We validate performative identity as we recognize that we don’t just “have” an identity, we perform it into existence through our daily linguistic choices and interactions. Furthermore, we build community by demonstrating how code-switching functions as a “we-code,” allowing us to establish solidarity to create a safe, inclusive environment for bilingual speakers.

Works Cited

Bahtina, D. (2026A). 6A: Conversation. UCLA Anthropology M150 Course Lectures.
Bahtina, D. (2026B). 8A: Practical work. UCLA Anthropology M150 Course Lectures.
Bahtina, D. (2026C). 7B: Multilingualism. UCLA Anthropology M150 Course Lectures.
Dingemanse, M., & Floyd, S. (2014). Conversation across cultures. In N. J. Enfield,
P.Kockelman, & J. Sidnell (Eds.), The Cambridge handbook of linguistic anthropology.
Cambridge University Press. https://doi.org/10.1017/CBO9781139342872.021
Duranti, A. (1997). Linguistic Anthropology. Cambridge University Press.
Gumperz, J. J. (1982). Discourse strategies. Cambridge University Press.
Auer, P. (2013). Code-switching in conversation: Language, interaction and identity. Routledge.
Zentella, A. C. (1997). Growing up bilingual: Puerto Rican children in New York. Blackwell.

Introduction and Background

Code-switching is a linguistic phenomenon that has consistently signaled community and group membership, in diasporas or otherwise. It is a feature of language that has often been used to show solidarity with other members of a marginalized group or to signal generational comradery. Persistent usages of code-switching have given rise not only to unique means of language usage, but to what some would consider entirely new languages. This can be seen not only through academic observations of bilingual speakers, but also through the lived experiences of those speakers. One particular space where both academic and personal reflections are abundant is regarding the usage of “Spanglish,” a unique combination of Spanish and English used by many in the Latinx community.

Previous research on Spanglish has found that in cities like New York City, Spanish-English bilingual youth have consistently reflected their cultural belonging and showcased their understanding of bilingual cultural norms through code-switching (Montes-Alcalá, 2024). These findings have helped build an understanding of how Spanglish is evoked as a cultural and social tool, rather than simply a linguistic phenomenon.

Personal narratives have also reinforced this perception of Spanglish as a cultural device. In speaking about her Spanglish usage during a Ted Talk, high school student Alondra Posada (2018) states that Spanglish helps people communicate, especially when they are struggling to learn both Spanish and English at the same time. She also states that Spanglish is a strong representation of both her American and Latinx communities, indicating that it helps her connect and feel comfortable within both elements of her identity. Spanish professor Dra Meduri Soto (2021) reinforces this concept when speaking about the U.S. Latinx community’s usage of Spanglish. She states that her relatives in Mexico usually admired Spanglish when they heard it spoken. She emphasizes that it was hard, if not impossible, to teach them Spanglish because she had acquired it organically as part of her American Latina identity. Generally, Spanglish speakers’ own conceptions and accounts of their Spanglish use reinforces its usefulness as a communication device that brokers understanding and a sense of cultural belonging.

Other studies have added onto this conversation by exemplifying how Spanish-English bilinguals have adapted their usage of Spanglish to other spheres, for example by using online content like memes to resonate with each other and with a broader Latinx audience (Dickinson, 2023). In the digital age, this usage of online resources has become increasingly significant, prompting one to wonder how community building via code-switching and Spanglish has been altered by social media apps and forums, as well as how digital connectedness has molded and shaped code-switching as a whole. These questions and qualms come at a time where content posted on Tik Tok specifically has taken the world by storm, allowing young people a new platform on which to gain popularity and connect with others. Existing research (Daulay et al, 2024) has already shown that code switching in Tik Tok and Instagram comments can shift subjects and showcase bilingualism between app users. Therefore, there is some basis to assume that platforms such as Tik Tok are already being used to create community and solidarity. However, not much is known, at present, about how specific linguistic strategies, such as intra-sentential switches or inter-sentential switches, function in TikTok’s humor-driven environment to build and sustain community connections. These key terms, as well as code switching as we use it in our research, have been defined in the following index:

This study seeks to fill the gap in knowledge regarding social media code-switching tools by analyzing Spanglish TikTok skits, focusing on Mexican and Puerto Rican dialects, to uncover how creators use language to entertain, unite, and resonate with bilingual audiences. By examining linguistic patterns and audience engagement, this research sheds light on how humor and language intersect to strengthen cultural identity in a rapidly evolving digital landscape. Ultimately, the research poses the question of exactly how Spanglish in Tik Tok skits is building humor and community in the Mexican and Puerto Rican online communities and how engagement metrics are reflective of this.

Before formally addressing this question, we had four personal hypotheses. For one, we believed inter-sentential switching would occur more often than intra-sentential switching. Next, we believed many videos would have both intra and inter-sentential switching. Third, we anticipated seeing many comments with personal pronouns and “humorous” emojis. Finally, we hypothesized that comments, likes, and shares would differ more than views when inter versus intra-sentential switching was used. These hypotheses were supported at varying degrees throughout our data collection process, as we will address.

Methods

To understand how bilingual people use Spanish and English in humorous TikTok videos, we picked eight content creators aged 18 and above who use skits and hashtags such as #Spanglish and #comedy. We watched eight videos where these creators mix languages, a practice we’ve already defined as code-switching. We counted how often they switched between languages and looked at two types of code-switches, both of which have also been defined above: intra-sentential switching and inter-sentential switching.

We studied how people responded to these particular videos by checking how many likes, comments, and shares each got. For example, we looked at what kinds of jokes people found funny, how many people saved the videos, and if they shared their own, often extremely similar, experiences in the comments. This helped us see which type of code-switching typically works best to make people laugh and relate to creators. We also noticed how creators used their language background, like “Spanish de rancho” and Puerto Rican Spanish, to connect with their audiences. By studying these videos, we aimed to learn how language blending can be a tool for creativity and connection.

Data Collected: Figure 1 showcases video analytics for a subset of the eight TikTok videos we analyzed for our results analysis. Figure 2 highlights types of code switching and examples of code switches utilized within the same videos. The five videos selected were the first five observed. Additionally, they included the videos with the highest numbers of code-switching, serving as valuable data to showcase.

Video Number	Dialects	Views	Likes	Comments	Top Comment Themes
1	“Spanish de Rancho” & American English	511.9k	68.5K	1283	1) Shared experience 2) Comedic effect, including mentions of laughter
2	Puerto Rican Spanish & American English	263.5k	67.3k	205	1) Shared experience
3	Puerto Rican Spanish & American English	92.4k	7626	101	1) Comedic effect, including usage of “cry” and “laugh” emojis
4	Mexican Spanish & American English	7.8M	1M	4599	1) Shared experience, particularly in familial settings
5	Mexican Spanish & American English	1.1M	156.5k	857	1) Shared experience 2) Comedic effect, including usage of “cry” and “laugh” emojis

Video Number	Dialects	Number & Types of Code-switches	Example of code-switch
1	“Spanish de Rancho” & American English	Total: 28 22 intra-sentential 6 inter-sentential	“I told my patient ‘encuerate’”
2	Puerto Rican Spanish & American English	Total: 5 3 intra-sentential 2 inter-sentential	“Estamos buscando una pillow”
3	Puerto Rican Spanish & American English	Total: 8 6 intra-sentential 2 inter-sentential	“I cannot live in a caliente place”
4	Mexican Spanish & American English	Total: 3 2 intra-sentential 1 inter-sentential	“How to say ‘to tag’ in spanish”
5	Mexican Spanish & American English	Total: 15 15 intra-sentential	“I speak spanglish of course cuando se me olvida..”

Tik Tok Analytics:

Tik Tok analytics were collected for various reasons. We wanted to see, more generally, how many views, likes, and comments Spanish-English creators were getting for their “Spanglish” centered skits and videos. More specifically, we wanted to be able to compare analytics between videos that used intra-sentential and inter-sentential switching to see if any one of them helped creators gain more popularity than the other. Within the eight total TikTok videos we observed, some words and ways of talking did seem to make videos more popular, garnering more likes and comments. It seemed as though solely using intra-sentential switching was less successful in pulling in likes, relative to the amount of views videos garnered. Figure 3, seen below, exemplifies this. As a note, the average ratios were calculated by taking the ratio of likes to views for each of the eight videos assessed and subsequently taking the mean amongst videos

As can be observed within the figure, videos with both types of switches were over two times as successful at gaining likes when compared to their counterparts with only intra-sentential switches. The videos with both inter and intra-sentential switching often used both types of switches to mirror instances between Spanish speaking parents and bilingual children, occasions which likely evoke relatability. On the other hand, the TikTok videos where people mix Spanish and English in the same sentence didn’t get as many likes or saves, perhaps because they were harder for the audience to follow.

A more qualitative analysis of the video comments also revealed that the Tik Tok audience often uses words in the comments like “I”, “myself”, and “me”, because they are relating to experiences showcased in the videos. They additionally use emojis like “😂” and “😭” to show laughter or emotion. Sometimes, the audience even mixes Spanish and English themselves.

Other Results: After assessing our eight selected Tik Tok videos, we found the following, displayed in Figure 4:

We borrowed methodology previously used in an analysis of code-switching in Instagram videos (Wiraputi et al, 2021). This method looks at not only the number of code switches, but also what percentage of code switches are of a certain type. The methodology we copied had “tag-switching” as well, an extra-sentential type of switching where a speaker inserts phrases from one language into another, but we did not notice the occurrence of tag-switching within our videos and omitted it as a result. The results that we did find showed a few interesting phenomena:

1. When Spanglish Tik Tok creators code switch in videos, they typically do so more than once. In fact, the average across eight videos was 9.25 code switches per video.
2. Intra-sentential switching dominates code switching occurrences, taking up over 77% of the code switching we witnessed.
3. Co-occurrence was in fact common. Only two of eight videos or 25% of content observed did not have both inter and intra-sentential switching.

Having analyzed all our quantitative and qualitative data on video analytics, comments, and occurrence of code-switches, our results yielded the following consequences as far as our hypotheses are concerned:

Discussion and Conclusion

This study of Spanglish humor on TikTok revealed that young adult creators frequently use intra-sentential code-switching, particularly in Mexican and Puerto Rican dialects, to enhance the comedic and relatable quality of their content. Contrary to our initial hypothesis, intra-sentential switching was more prevalent than inter-sentential switching, indicating that the integration of Spanish and English within sentences may occur more commonly with bilingual audiences. Another key insight we came across was the ways in which community and identity building proceeds amongst the Latinx community in the United States. Prior to our studies, we had understood that Spanish varieties, such as that spoken in Los Angeles, are often characterized by convergence, borrowings, and switches (Sánchez-Muñoz, 2017). We had been curious about the role that convergence played on social media and on Tik Tok specifically and whether the blends we witness in real life translate to online spheres. Our analysis pointed us towards the observation that Spanglish serves as a vital tool for expressing cultural identity and fostering community solidarity online, making humor a bridge between bicultural experiences. We also noticed while studying engagement patterns that videos that incorporate both intra and inter-sentential switching see higher engagement, pointing to the idea that utilizing many different linguistic patterns within the same video boosts performance.

Some considerations we have for future research into this field is the generalizability of our findings. Since our findings are specific to TikTok and its user demographics, they may not apply universally across all digital platforms or linguistic communities. Comparative studies across different digital platforms and research completed over a longer period of time could explain how digital code-switching presents itself in other contexts. In essence, the study underscores the importance of linguistic flexibility in digital content, which not only entertains but also strengthens community ties among bilingual Latinx viewers.

References

Dickinson, K. V. (2023). What Does It Meme? English–Spanish Codeswitching and Enregisterment in Virtual Social Space. Languages, 8. https://www.researchgate.net/publication/374657133_What_Does_It_Meme_English-Spanish_Codeswitching_and_Enregisterment_in_Virtual_Social_Space

Hamidah Daulay, S., Husein Nst, A., Randia Ningsih, F., & Berutu, H. (2024). Code Switching in the Social Media Era: A Linguistic Analysis of Instagram and TikTok Users. Humanitatis Journal of Language and Literature, 10. https://www.researchgate.net/publication/383171193_Code_Switching_in_the_Social_Media_Era_A_Linguistic_Analysis_of_Instagram_and_TikTok_Users

Montes-Alcalá, C. (2024, April 15). Bilingual texting in the age of emoji: Spanish–English code-switching in SMS. MDPI. https://www.mdpi.com/2226-471X/9/4/144

Sanchez-Munoz, A. (2017). Tempted by the Words of Another: Linguistic Choices of Chicanas/os and Other Latinas/os in Los Angeles. https://www.researchgate.net/publication/345124311_Tempted_by_the_Words_of_Another_Linguistic_Choices_of_Chicanasos_and_Other_Latinasos_in_Los_Angeles

Soto, Dra. M. (2021, September 15). We Speak Spanglish ¿Y qué? World Outspoken. https://www.worldoutspoken.com/articles-blog/spanglish-its-who-we-are

Sanchez-Munoz, A. (2017). Tempted by the Words of Another:: Linguistic Choices of Chicanas/os and Other Latinas/os in Los Angeles. https://www.researchgate.net/publication/345124311_Tempted_by_the_Words_of_Another_Linguistic_Choices_of_Chicanasos_and_Other_Latinasos_in_Los_Angeles

TEDx Talks. (2018). Spanglish is a Language Too! | Alondra Posada | TEDxYouth@UrsulineAcademy. In YouTube. https://www.youtube.com/watch?v=6N8Zk_EY6G8

Wiraputri, D., & Sultara, K. (2021). Code switching found in cinta laura’s video on her instagram tv. Journal of Language and Applied Linguistics.

Introduction and Background

In today’s world, bilingual music is a new and exciting way for artists to express their identity and connect with different groups of people. Look at Rosalía or Kali Uchis, for example: they both switch seamlessly from Spanish to English, often in the middle of their lyrics. What’s really happening when these artists switch languages so effortlessly? Is it just a trendy way to mix things up, or is more happening beneath the surface?

This blog delves into how bilingual artists navigate the linguistic restrictions surrounding code-switching in music. We will test if certain constraints governing spoken Spanish and English also apply in a musical context. Specifically, we will look at whether Johanson’s (2011) code copying framework and Poplack’s constraints (1980) still hold up in a lyrical format.

To understand what happens when bilingual artists code-switch in their lyrics, it helps first to unpack what code-switching is. Code-switching occurs when a speaker alternates between two languages within or between sentences. For example, Kali Uchis has a song lyric that goes, “That’s the only thing que me da alegría,” which means “That’s the only thing that brings me joy.” Here, she casually switches from English to Spanish mid-song.

According to Poplack (1980), there are some crucial grammatical constraints that bilinguals tend to follow when switching between languages, including the free morpheme constraint, which suggests that bilinguals are more likely to switch languages when one of the languages is a free morpheme. This helps to ensure that the switch does not disrupt the sentence’s grammatical structure. However, Timm (1975) argues that not all code-switching follows strict rules. Some switches and combinations of words may sound ungrammatical to bilingual speakers. Timm states that bilinguals make code-switching decisions based on the context and what feels natural in the conversation. In other words, code-switching is as much about communication goals and cultural expression as it is about grammar. However, there is another layer to all this, which is that of music. Lyrics are more than just spoken language; they are a form of artistic expression, so they have more leeway to break the “rules” of language and grammar regarding creativity. In music, artists like Rosalía and Kali Uchis often code-switch between English and Spanish to communicate and express their identity, emotions, and cultural background. In this context, code-switching may not always conform to the strict linguistic rules established by Poplack and others. These two artists’ flexibility in language reflects their different bilingual experiences and how they shape how they mesh languages in their songs.

Methods

To understand and answer our research question, we analyzed the lyrics of two prominent Spanish-English artists, Rosalía and Kali Uchis. We chose these two artists for their unique linguistic and cultural backgrounds and active use of code- and language-switching in their music. In terms of similarities, both artists are of a similar age and release Latin/Spanish music of similar genres. We thought this genre of music in particular, would be insightful to focus on, as the informality and structure of their song lyrics are a contrast from previous code-switching examples we have looked at through this course. Both women have also each built an alternative, trend-setting brand image and self-identities that hinge largely on themes of confidence, sex positivity, femininity, and empowerment.

Rosalia, photo via Europa Press News                            Kali Uchis, photo via Felipe Q Noguiera

Beyond these similarities that allow for standardization of our data collection, we chose Rosalía and Kali Uchis because they also represent two distinct types of bilingualism and cultural experiences, providing a rich contrast for analysis. Rosalía, born in Catalonia, Spain, is a sequential bilingual who grew up speaking Spanish and Catalan as her native languages before learning English later in life — we categorized her as an L2 English speaker. While she began her music career with a strong focus on Flamenco and traditional Spanish music, Rosalía recently transitioned into Latin and alternative reggaetón-inspired pop. Kali Uchis, in contrast, is a simultaneous bilingual born in the United States to Colombian parents. Growing up, she was exposed to both Spanish and English equally, having spent significant time in both Virginia, USA and Colombia. This dual-culture childhood is reflected in the blending of the two languages in her songs. Her discography began predominantly in English but has recently shifted to more Latin music. By examining these two artists, we hoped to explore how their bilingual backgrounds shape their distinct approaches to code-switching and linguistic creativity and how their different upbringings may have manifested themselves into their music.

We collected data by identifying ten songs from each of these two artists. We used the metric of the number of Spotify streams per song to choose their most popular music. In our data filtering, we only ensured to include tracks where the artists held writing credits, as this guaranteed that the observed code-switching and linguistic patterns came directly from their creative process rather than external songwriters that we would not know their linguistics backgrounds. Because of this in mind, we chose to include collaborations and feature songs, such as New Woman (LISA feat. Rosalía) and SI NO ES CONTIGO – REMIX (Cris Mj feat. Kali Uchis & JHAYCO), as these song features tended to be 1) highly streamed, and 2) written by the artists themselves. Then, to narrow our focus to bilingualism in music, we excluded songs entirely in Spanish or English, as these songs were not tailored to our interests and would not illustrate instances of bilingual code-switching. Lyrics were sourced from Genius, a widely used database for accurate song transcription, translations from the Genius community comments, or from members of our group who are beginning L2 Spanish speakers.

Example of songs taken from Rosalia’s Spotify artist profile, including streams

After selecting 23 songs for our dataset, we parsed each song, compiling lines of lyrics in which a language switch occurred. To identify these switches, we applied the established definitions of intra- and inter-sentential switching and the structure of the code-copying framework. These lyrics were documented and later analyzed on a robust spreadsheet.

Song Selection Data Samples
Kali Uchis
Song: SAD GIRLZ LUV MONEY Remix – Feature   2021 Release 461.5m Approx. Spotify Streams	Lyric: Yeah, you’ve been starin’ at me (¿Por qué me miras?) That’s the only thing que me da alegría Yo quiero sentirte inside of me
Song: Moonlight   2023 Release 829.8m Approx. Spotify Streams	Lyric: I just wanna get high with my lover, Veo una muñeca cuando miro en el espejo
Song: telepatía   2020 Release	Lyric: If you want it, you could take a private plane, a kilómetros estamos conectando

1.2b Approx. Spotify Streams	Y me prendes aunque no me estés tocando
Song: Igual Que Un Angel   2024 Release 332m Approx. Spotify Streams	Lyric: You should’ve seen the way she looked, igual que un angel Heaven’s her residence y ella no se va a caer
Rosalia
Song: Despecha   2022 Release 1.1b Approx. Spotify Streams	Lyric: Hoy salgo con mi baby de la disco corona, corona, yeah
Song: New Woman – Feature   2024 Release 223m Approx. Spotify Streams	Lyric: Puta, soy la Rosalía, solo sé servir
Song: TKN   2020 Release 465m Approx. Spotify Streams	Lyric: Zoom en la cara, Gaspar Noe
Song: Besos Moja2   2022 Release 442.6m Approx. Spotify Streams	Lyric: To’ me queda bien con una white tee To’ me quieren casar, soy una wifey

Results and Analysis

In total, we collected data from 23 different songs released between 2019 and 2024, with 10 being songs by or featuring Rosalía and 13 by or featuring Kali Uchis. From these 23 songs, we extracted 19 (38.7% of data) data samples from Rosalía and 30 (61%) data samples from Kali Uchis – resulting in 49 data samples. Generally, most code-switching occurs as an intersentential switch, which means it happens at clausal boundaries between sentences instead of within a sentence.

The results we discovered while gathering our data prove some interesting differences between how sequential and simultaneous bilinguals code-switch. For example, Rosalía, a sequential bilingual who primarily makes her music in Spanish within the Flamenco musical style, is more likely to code switch in her music when in the Latin context or if she appears as a feature on songs with Latin artists. Meanwhile, Kali Uchis, a simultaneous bilingual, is more likely to code-switch in her music in general. Unfortunately, due to time constraints, we could not investigate how heavy of a correlation each artist’s background influences their code-switching/code-coping tendencies. Still, based on the data, we can tell each artist has a preference on how bilingualism is written into their music. We expected a vast majority of code copying to occur within our data. Still, only 40.8% of data resulted from code-copying versus code-switching, with an overwhelming majority of code-copying data samples from Rosalía (80% of code-copying samples). Comparatively, Kali Uchis heavily favors code-switching, with 80% of her data samples containing some form of code-switching. She tends to favor longer phrases of code-switching instead of single-word copies. The data table below shows examples of the different types of code-copying occurring with a song from each artist.

Song Lyric Data Analysis Samples
Igual Que Un Angel – Kali Uchis Kali Uchis, Peso Pluma – Igual Que Un Ángel 0:20-0:30
Lyric: You should’ve seen the way she looked, igual que un angel Heaven’s her residence y ella no se va a caer	Translation: You should’ve seen the way she looked like an angel Heaven’s her residence and she’s not gonna fall	Analysis: Selective copying: igual que un angel mirrors the English structure to retain meaning while adapting to Spanish syntax
fue mejor – Kali Uchis Kali Uchis – fue mejor feat. SZA 1:00-1:10
Lyric: Y me fui en el Jeep a las doce The backseat donde yo te conocí	Translation: And I left in the Jeep at twelve The backseat where I met you	Analysis: Global copy example: “Jeep” and “the backseat” are English words with all properties preserved and integrated into the lyrics
Besos Moja2 – Rosalia

Wisin & Yandel, ROSALÍA – Besos Moja2 2:22-2:33
Lyric: To’ me queda bien con una white tee To’ me quieren casar, soy una wifey	Translation: It looks good on me with a white tee They all want to marry me, I’m a wifey	Analysis: Global copying: the English phrases white tee and wifey are borrowed intact and function within the Spanish sentences without alteration
New Woman – Lisa feat. Rosalia LISA – NEW WOMAN feat. Rosalía 1:48 – 1:52
Lyric: Puta, soy la Rosalía, solo sé servir	Translation: Bitch, I’m Rosalía, all I do is serve	Analysis: Selective copy example: semantic meaning of “serve” is preserved, directly translated from English to Spanish

The types of copying that we came across in our data are:

• Global copies (taking words in their unaltered form and using them within a sentence).
• Selective copies (the structure of a word or phrase is kept while adapting to another language’s syntax).
• Mixed copies (switching between one language to another within a sentence or phrase). From our data, a majority of the data uses mixed copying, accounting for 40% of the examples. Global copying accounted for 36% of our data, while selective copying accounted for 16%.

Discussion and Conclusion

From our corpus of lyrics, we concluded that code-switching in music tends to deviate from standard grammatical constraints and, therefore, does not strictly follow Poplack’s Constraints. Instead, code-switching in music more closely aligns with and reflects the artists’ bilingual background. Rosalía, a sequential bilingual, is more deliberate in her switches when using English in songs to identify specific concepts like “wifey” or “to serve.” On the other hand, Kali Uchis, a simultaneous bilingual, is more spontaneous in her use of code-switching; her fluidity between Spanish and English is representative of the dual integration of both languages in her identity. Still, we observed both global and selective copying in the artists’ lyrics and could conclude that global and selective copying are common and focal aspects of bilingual lyricism.

Nonetheless, we acknowledge the limitations of our investigation. Due to the temporal constraints of the quarter system, our project timeline was very restricted, and we could not expand the scope of our exploration and means of data collection. Thus, our focus on two artists and reliance on public information could have been a more comprehensive elicitation of analyses.

Revisiting our initial research questions: If certain constraints apply in spoken Spanish when code-switching to English, do they transfer over the same way into language in a musical context? What types of copying, following the code copying framework, will be found in song lyrics?

The integration of multiple languages in music has far-reaching implications for linguistic research, cultural understanding, and artistic expression. Understanding code-switching in creative language use provides insights into how language patterns manifest in nontraditional language contexts, namely under the influence of the artistic nature of lyricism and songwriting. Code-switching in music also enhances artistic expression by encouraging dynamic wordplay and rhymes across languages. It provides insights into the differences between different levels of bilingualism and types of multilingualism. Moreover, bilingual lyrics serve important cultural functions, as they represent the juxtaposition of cultures, promoting linguistic and cultural diversity, expressions, and communication. Within the music industry, multilingual music can unite different listeners’ cultures and expand the reach of artist audiences.

Although our research yields valuable insights, it needs to be more broadly generalizable to all Spanish-English or other language bilingual musicians. Future expansions of our study will include a larger sample of bilingual artists with a more extensive range of lyrical repertories from more diverse language backgrounds. It would be ideal to have direct access to artists’ accounts of their language to better understand their use and relationship with their respective languages as well as their creative processes.

An important consideration to account for in our investigation is the existence of code-switching patterns for Spanglish that do not necessarily apply to varieties of Iberian Spanish. The present investigation focused primarily on code-switching conventions within a musical genre rather than a dialect, assuming that an artist from Spain would adapt to Latin American dialectal features in this context, but dialectal differences may still occur. We are interested in understanding competing factors of bilingual acquisition and regional norms to further develop our research from this investigation, specifically, the influences of sequential and simultaneous bilingualism and regional norms for language in Latin American versus peninsular Spanish varieties. Additional research directions would look across different genres and investigate the reception of code-switched lyrics among monolingual and bilingual listeners, considering the linguistic influence of lyricism more broadly in language communities.

References

Cris MJ, Kali Uchis & JHAYCO – Si No Es Contigo (remix). Genius. (n.d.-a). https://genius.com/Cris-mj-kali-uchis-and-jhayco-si-no-es-contigo-remix-lyrics

Duran, R. P. (1984). Latino Language and Communicative Behavior. The Modern Language Journal. Exposito, Suzy. “Pop Loner Kali Uchis on Growing up Punk in Colombia and the Struggle to Stay Bilingual.” Rolling Stone, Rolling Stone, 13 Sept. 2019, www.rollingstone.com/music/music-latin/kali-uchis-isolation-colombia-interview-2018-7 00772/.

Gonzalez-Cruz, M. I. (2017). Exploring the dynamics of English/Spanish codeswitching in a written corpus. Alicante Journal of English Studies, 30 (2017), 336-361.

Johanson, L. (2011). Contact-induced change in a code-copying framework.

Lisa (ft. Rosalía) – New Woman. Genius. (n.d.). https://genius.com/Lisa-new-woman-lyrics Lipski, J. M. (2008). Varieties of Spanish in the United States. Georgetown University Press.

Loaiza, K. M. “Kali Uchis”. Lyrics to “ORQUÍDEAS” [Album]. Geffen Records, 2024. Genius, genius.com/albums/Kali-uchis/Orquideas.

Monteagudo, M. (2020). Spanglish code-switching in Latin pop music: functions of English and audience reception. https://www.duo.uio.no/handle/10852/79796

Poplack, S. (1980). Sometimes, I’ll start a sentence in Spanish Y TERMINO EN ESPAÑOL: toward a typology of code-switching¹. Linguistics, 18(7-8), 581-618. https://doi.org/10.1515/ling.1980.18.7-8.581

Sayre, A. A. (2022). “Rosalia is unafraid to pull from every corner of the world.” NPR, NPR, 24 April 2022, https://www.npr.org/2022/04/24/1094491641/rosalia-is-unafraid-to-pull-from-every-corn er-of-the-world

Timm, L. A. (1975). Spanish-English Code-Switching: El Porqué y How-Not-To. Romance Philology, 28(4), 473–482. http://www.jstor.org/stable/44941606

Villa Tobella, R. ‘ROSALÍA’. Lyrics to “Motomami” [Album]. Columbia Records and Sony Music, 2022. Genius, genius.com/albums/Rosalia/Motomami.

Rosalía. (2022). Motomami [Album]. Columbia Records. Uchis, K. (2024). Orquideas [Album]. Geffen Records. Uchis, K, and SZA. “Fue Mejor.” Genius, Genius Media Group, n.d., https://genius.com/Kali-uchis-and-sza-fue-mejor-lyrics

In this article, we investigate how code-switching in K-pop functions, focusing on its role in enhancing audience engagement. Intra-sentential is a type of code-switching where elements from one language (English) are embedded within a sentence structured primarily in another language (Korean). For example, a K-pop song might use an English verb or noun phrase, such as “make it rain” or “crazy 사랑” (crazy love), within a Korean sentence. This use of English within Korean sentences can enhance the song’s appeal to both Korean listeners and a global audience. Our target population includes individuals between the ages of 18 and 21, both L1 Korean monolingual speakers as well as global listeners, L1 English monolinguals, with minimal K-pop exposure. Our research will analyze phonological adaptation focusing on rhythmic and prosodic adaptations of English within Korean phrasing. This type of adaptation involves adjusting English words to fit the rhythmic and prosodic structure of Korean music. K-pop blends English into Korean lyrics by adapting to differences in rhythm where Korean’s syllable timing versus English’s stressed pattern makes the music appeal to both global and Korean listeners. The theoretical framework we will use is the Myers-Scotton’s Matrix Language Frame (MLF) to analyze how English integrates into the Korean structure and appeals to monolingual Korean listeners. The selected artists are global K-pop stars popular in LA, an ideal setting to study phonological adaptability due to the growing local fanbase of the genre (Burt, 2024). Our research question is: How has English usage in K-pop evolved across different popularity levels, and how does this linguistic adaptation resonate with different subsets of audiences (L1 Korean monolinguals versus global appeal)? Our hypothesis suggests that the use of English in K-pop has grown between 2015 and 2024. As its popularity grew internationally, code-switching has increased between Korean and English. In songs aimed at global audiences or informal settings, we expect more frequent code-switching and phonological adaptations. In contrast, songs focusing on cultural authenticity or Korean listeners may use English more selectively. Due to strategic linguistic adaptations, the patterns of syllable and stress have evolved in response to the increasing globalization of K-pop. Before this study, the specific patterns of phonological adaptation of English within K-pop and their differential impact on L1 Korean versus L1 English listeners were not fully understood.

Methods

Before looking straight into the phonological particles of this study, we want to understand the difference between the English language of older and newer K-pop songs. To do so, two participating groups completed a demographic Google survey sent through emails: Three L1 English monolinguals with minimal K-pop exposure and three L1 Korean monolinguals. These two groups were explicitly chosen because L1 Korean monolinguals understand and speak the domestic language of the songs(Korean), and L1 English monolinguals with minimal exposure have limited background knowledge of the Korean language, representing the global language and audience.First participants were asked general demographics, then presented with some terms and concepts to help participants understand the study better, such as rhythm and stress and what older era, songs released before 2015 and newer eras were songs released after 2015. After understanding the terminology, they were told to listen to the first two verses of two songs: one older and one newer release track from one of the K-pop artists, BTS and Seventeen, who are popular in LA. A little background was that the debut of PSY’s song “Gangnam Style” paved the way for international concerts to rise, with the global market growing by 44.8% in 2020.(Glasby, 2021; Oos, 2023). The years 2015 and 2022 were chosen with this fact in mind, where in 2015, many of the current popular K-pop artists debuted. One of them is Seventeen, and the other is BTS. The songs from Seventeen were: “Mansae” (2015) and “HOT” (2022), and the songs from BTS were “DOPE” (2015) and “Yet to Come” (2022). At the end of the survey, we asked two essential questions: 1) Do the English lyrics in this song match the Korean rhythm and pronunciation, and 2)Can you identify specific words or phrases where the English lyrics were adapted? By identifying the two questions, we analyzed the differences in which they noticed and what type of changes each group noticed more.

In order to conduct our qualitative data, we investigated stress patterns in the English lyrics of K-pop songs using Praat software. Stress was measured by analyzing spectrograms and waveform data, identifying syllables with higher intensity peaks, and recording their durations in milliseconds. These measurements helped us determine how stress in English lyrics was adapted to fit the balanced rhythmic structure of Korean songs. In this analysis, we also examined the alignment of stressed syllables with surrounding Korean lyrics to understand the degree of rhythmic congruence.

The methodology was designed to capture the differences in stress and rhythm between Old and New Era songs, using precise timestamps to compare English syllables embedded in bilingual lyrics. For example, in “Dope” (BTS), an Old Era track, the word “party” maintained its natural English stress pattern, with the second syllable emphasized and lasting approximately 200 milliseconds longer than the average Korean syllable. By contrast, in “Yet to Come” (BTS), a New Era song, the phrase “dreams come true” exhibited more balanced syllable durations, where each syllable was approximately 140 milliseconds, blending seamlessly into the Korean rhythm. These observations were validated through auditory analysis to ensure that the measurements aligned with listeners’ perception of rhythmic congruence.

Exploring stress and rhythmic patterns in this way is critical for understanding bilingual media consumption. In bilingual contexts like K-pop, successful integration of English lyrics into Korean songs depends on aligning distinct rhythmic and phonological systems. By adapting stress patterns to conform to Korean syllable timing, producers create a more cohesive listening experience for both local and global audiences. This analysis thus sheds light on how bilingual media can balance cultural authenticity with accessibility, enhancing its appeal across linguistic boundaries. To collect quantitative data, once again using Praat, we looked into one line from each song that contained intra-sentential code-switching to analyze the syllable duration for each language. Separating each syllable into its own section, we first measured the syllables in milliseconds and added each syllable to its belonging words. This is due to the need to account for the rhythmic adaptation of English stress patterns within Korean phrases, keeping both languages’ syllable duration more precise to compare. Then, we calculated the average syllable duration for each language and compared the length in older and newer songs to see if, in newer songs, the English words were phonologically adapted to align with their Korean counterparts.

Results and Analysis

Figure 1: BTS “DOPE” syllable duration Results and Analysis

Our study revealed increased integration of English lyrics and phonological adaptations over time. This was first highlighted in our findings from our Google survey, the data of which was converted into a graph generated by Python.

The graph below represents two pieces of data collected from the Google form. The left graph shows the results from “English Words Noticed in Older vs. Newer Songs,” and the right graph shows the “Adaptation Noticed in Older vs. Newer Songs.”

Figure 2: Google Survey Python Generated Charts

The first graph revealed that L1 English monolinguals with minimal K-pop exposure noticed significantly more English words in newer songs than in older songs. In older songs, they noticed a mean of 5 words and a mean of 15 in newer songs. Similarly, L1 Korean monolinguals also saw an increase in English lyrics, though they noticed slightly less than L1 English monolinguals. In older songs, they noticed a mean of 4 and 10 words in the newer songs. In contrast, the second graph revealed that L1 korean monolinguals were more likely to notice phonological adaptations in newer songs, reflecting shifts toward a more balanced syllable usage that aligns English with Korean rhythm and timing. As presented in the graphs, L1 English speakers were slightly less likely to realize these phonological adaptations. These findings underline how newer K-pop tracks cater to both global and domestic audiences through strategic adaptations where English lyrics are adjusted to preserve their clarity while fitting into the korean rhythm.

Our study highlights a significant evolution in the integration of English lyrics and their phonological adaptations in K-pop over time. Using Praat software, we observed how syllable durations evolved, showcasing the interplay between English stress patterns and Korean rhythmic norms across Old and New Era songs.

In “Dope” (BTS), an Old Era track, English stress often disrupted Korean rhythm. For instance, the syllable “Gent-” in “Ladies and Gentlemen” lasted 200 ms, longer than the Korean syllable “Ayo” at 180 ms. Similarly, “Cause we got fire” featured “fire” at 330 ms, highlighting mismatched timing. Conversely, the New Era track “Yet to Come” (BTS) exhibited greater alignment, with “past” and “best” balanced at 690 ms, closely matching adjacent Korean syllables. A similar trend was seen in Seventeen’s “Mansae”, where mismatches occurred with “name” at 310 ms and “teen” at 280 ms. However, in “Hot”, the New Era song, phrases like “Spicy feelin’” featured evenly timed syllables (~210 ms), harmonizing with the Korean rhythm. Aggregated averages reinforced these findings: in “Dope”, Korean syllables averaged 535 ms, while English averaged 1,183 ms, creating a 500 ms gap. This narrowed significantly in “Yet to
Come” to a 100 ms gap (639 ms vs. 726 ms). Similar alignment improvements were observed in Seventeen’s tracks. These results highlight the significance of adapting English stress patterns to fit Korean rhythms in bilingual media consumption. Such adaptations enhance accessibility for both global and local audiences, enabling K-pop to transcend language barriers. By harmonizing distinct phonological systems, producers create music that resonates culturally and emotionally with diverse audiences. This strategic linguistic integration underscores K-pop’s role as a global phenomenon, balancing cultural authenticity with international appeal.

Figure 3: Syllable Duration of Old Vs. New Era Songs Generated via Python

Figure 4: Syllable Duration for Phrases of Interest (Old vs. New Era) Generated via Python

Figure 5: Python Code For Figures 3 & 4

With the praat software, we calculated the syllable duration, looked into the average syllable duration, and input it into a bar graph using Google Sheets. In the first Figure, we analyzed the syllable duration in the signs between Old and new-era songs from BTS

Figure 6: BTS “DOPE” and “Yet to Come” Syllable Duration Comparison

In the 2015 song “DOPE”, we saw a significant gap between the average syllable duration in the korean lyrics, n=535ms, and the English lyrics, n=1,183ms, resulting in a 500ms gap. But when we looked at BTS’s new-era song “Yet to Come” in 2022, the gap between the average syllable duration for Korean, n=639ms, and English, n=726ms, the duration was closer with about a 100ms gab, aligning each syllable duration closer. The same case was shown in the artist Seventeens old-era song “Mansae” 2015 and the new-era song “HOT” 2022.

Figure 7: Seventeen “Mansae” and “HOT” Syllable Duration Comparison

In “Mansae,” the average Korean Syllable duration was n=400, and the English syllable duration was n=215, resulting in a 200 ms difference. Comparing this to the newer song “HOT,” the Average Korean Syllable duration was n=237ms, and the English syllable duration was n=309ms, a 100 ms difference. Though Seventeen’s results showed less significant change than those of BTS, both cases displayed a decrease in the duration of Korean and English lyrics in their old-era and new-era songs. BTS’s songs reflected a decline in English syllable durations compared to their newer songs; in contrast, Seventeen songs demonstrate an increase in English syllable durations. Both show how the English syllables were adjusted to align with the korean syllable durations

Discussion and Conclusion

K-pop strategically uses code-switching and phonological adaptations to engage diverse audiences while maintaining its cultural identity. Despite increased English usage due to globalization, K-pop adapts English lyrics to fit Korean phonological patterns, resonating with domestic and global fans. This reflects a careful balance between cultural preservation and international reach. L1 Korean and L1 English listeners showed increased recognition of English lyrics in newer K-pop songs. However, for different reasons, L1 English listeners noted the greater prominence of English, while L1 Korean listeners recognized the phonological adaptations to Korean rhythms. This suggests intentional design to appeal to both groups, reflecting either listener adaptation or producer strategies to bridge linguistic divides.

While globalization has amplified English use, K-pop prioritizes cultural identity. Newer songs like BTS’s “Yet to Come” and Seventeen’s “HOT” demonstrate smoother integration of English lyrics into Korean rhythms compared to older tracks like “Dope” and “Mansae,” ensuring a cohesive listening experience for Korean audiences while making English more accessible to international fans. K-pop thus acts as a bridge between local traditions and global influences, transcending language barriers without losing its roots. Higher English lyric recognition in newer songs underscores deliberate linguistic strategies by K-pop producers, highlighting the genre’s ability to adapt to globalization while preserving its cultural identity. Further research is needed to determine the consistency of these trends across different K-pop groups and genres, and expanding participant diversity could provide deeper insights into how linguistic adaptations influence listener engagement and cater to domestic and international audiences.

This study illustrates how K-pop reflects broader globalization in the media. By embracing English through linguistic adaptation, K-pop reinforces cultural identity while connecting with global audiences. Balancing cultural authenticity with international appeal, K-pop offers a model for navigating bilingualism and globalization. It contributes to discussions about language’s role in cultural preservation and global expansion and demonstrates how media can connect audiences worldwide while maintaining unique cultural heritage.

References

Burt, K. (2024). How LA became the epicenter of K-Pop fandom in the U.S. Thrillist. https://www.thrillist.com/entertainment/los-angeles/k-pop-in-la

Can Code-Switched texts activate a knowledge switch in LLMs? A case study on English-Korean Code-Switching. (n.d.). https://arxiv.org/html/2410.18436v1

Canagarajah, S. (2007). Lingua Franca English, multilingual communities, and language acquisition. Modern Language Journal, 91(s1), 923–939. https://doi.org/10.1111/j.1540-4781.2007.00678.x

Glasby, T. (2021, May 19). How K-Pop took over America: A timeline. Nylon. https://www.nylon.com/entertainment/timeline-of-k-pop-rise-in-america

Le, S., Campbell L. J., Orozco J., & SS, N. (2024). Code-Switching Habits of BTS. https://languagedlife.humspace.ucla.edu/bilingualism/code-switching-habits-of-bts/

Lee, J. (2020). Cross-linguistic prosody in K-pop music: Integrating English and Korean lyrics. Journal of Music and Linguistics, 12(3), 40–55. Luminate Data (2023). Mapping out K-pop’s global dominance. Retrieved from Luminate Data’s blog on global streaming trends.

Nazri, S. N. A., & Kassim, A. (2023). Issues and Functions of Code-switching in Studies on Popular Culture: A Systematic Literature review. International Journal of Language Education and Applied Linguistics, 13(2), 7–18. https://doi.org/10.15282/ijleal.v13i2.9585

Oos, O. (2023b, December 5). When did K-Pop start becoming a global phenomenon – ononestudios.com. ononestudios.com. =The%20genre%20began%20to%20gain,to%20television%20dramas%20and%20cuisine

Rothman, J. (2009). Understanding the nature and outcomes of early bilingualism: Romance languages as heritage languages. International Journal of Bilingualism, 13(2), 155–163. https://doi.org/10.1177/1367006909339814

Schneider, I. (2023). English’s expanding linguistic foothold in K-pop lyrics: A mixed methods approach. English Today, 40(2), 105-112. https://doi.org/10.1017/S0266078423000275

Shim, D. (2005). Hybridity and the rise of Korean popular culture in Asia. Media Culture & Society, 28(1), 25–44. https://www.researchgate.net/publication/254737351_Hybridity_and_the_Rise_of_Korea n_Popular_Culture_in_Asia

Zhang, J. (2023). Language, cultural hybridity, and resistance in K-Pop: A linguistic analysis of Korean pop music lyrics and performances. International Journal of Emerging Multidisciplinary Business Information Systems, 3(3), 1-15. https://doi.org/10.59889/ijembis.v3i3.186