1. Introduction
The Hispanic or Latino community has become the second-largest ethnic group in the U.S., representing 18.7% of the total population (U.S. Census Bureau, 2020). Some U.S. geographic regions, such as the Southwest and Northeast, are home to long-standing communities with several generations of Latinos, whereas others, like the Southeast–which includes the state of North Carolina–are in earlier phases of development. According to the NC State Demographer (n.d.), this community will represent 11.3% of the total population of North Carolina (NC) in 2024. Due to the rapidly growing Hispanic population in NC, Michnowicz and colleagues (2023) emphasize that the Southeastern United States, and more specifically NC, is an important region to study. Due to the significant increase in the Hispanic population over the past three decades, they argue that unlike in other regions with long-standing Hispanic and Latino communities, NC presents the “opportunity to examine the initial stages of language and dialect contact as they emerge in real time” (p. 1).
Michnowicz and colleagues (2023), discussing Spanish in NC in particular, indicate that the observed characteristics in this community reflect a bilingual and bicultural identity in a new and evolving bilingual context, influenced by the growing Latino population in the region. Despite the increasing number of studies in recent years about the use and evolution of Spanish that examine language contact phenomena in NC (e.g., Carter, 2005; Michnowicz et al., 2018; Michnowicz, Trawick, et al., 2023; Ronquest et al., 2020) current research has failed to directly address the relationship between identity and contact phenomena. The current study focuses on bilingual heritage speakers belonging to the second (G2) and third G3) generations, who can be defined as bilinguals who “learned a language at home that is not the dominant language of the country” (Aalberse et al., 2019, p. 1).[1]
Research on various communities throughout the United States has indicated that first-generation immigrants (G1) maintain ties to their countries of origin through the retention of their home language, which decreases among subsequent generations (Escobar & Potowski, 2015; Michnowicz, Trawick, et al., 2023; Rumbaut, 2004; Shabtaev et al., 2022). Additional research reports similar effects regarding ethnic self-identification (hereafter ESI), which tends to shift from identification with national origin to identifying with the country of birth (e.g., the United States) among newer generations (Geerlings et al., 2015; Rumbaut, 2004). How, then, do Latinos residing in this newly forming community self-identify? That is, which ESI labels do they employ, how do they conceptualize ESI labels in different contexts, and how might they differ from Latinos residing in other more well-established bilingual communities?
Finally, in situations of language contact, a common phenomenon that often emerges among bilinguals is code-switching (hereafter, CS), which is defined as the use of two linguistic systems within “the same conversational event” (Toribio, 2011, p. 89). While early accounts of CS often described it as evidence of a lack of linguistic proficiency, considerable work has established that CS is an indicator of high levels of bilingualism in addition to serving as a marker of a dual identity (Myers-Scotton, 1995).
The primary goal of the present study is to investigate key aspects of sociolinguistic identity among G2 and G3 Hispanics of Mexican descent residing in North Carolina. We explore: 1) which ESI labels individuals select within and outside of the local community, along with their perspectives on these labels; 2) their preferred or chosen language (Spanish or English) in various social contexts; 3) the types of CS they exhibit; and 4) the relationship between their ESI labels, language choice, and the types of CS observed in their speech.
2. Literature Review
2.1. Language and Identity
Human identity is complex and flexible, changing according to the different roles individuals represent. Language serves as a crucial element in balancing these roles, as the use of a particular linguistic form or adherence to particular speech norms is one of the primary means of expressing our identities, including ethnic identities (Fought, 2006). As Bailey (2010) observes, “[l]anguage is our primary symbolic tool for representing and negotiating social reality, and it plays a central role in these processes of self and other-ascription […] We sometimes claim identities explicitly […] but much more often we perform them by speaking” (p. 76).
Fought (2006) indicates that scholars from various disciplines agree that ethnicity, much like race, is a social construct without an objectively measurable criterion; however, both remain real and crucial terms as they contribute to the creation of social reality. Additionally, ethnicity is relevant in multicultural societies as it involves “a complex process of constructing and reproducing identities within a particular community, a process intertwined with social, historical, ideological, and biographical factors” (p. 16).
Ethnic identity and language are intimately linked, as the way language is used affects how ethnic identity is shaped, and in return, ethnic identity influences the patterns of ethnic language utilization (Giles & Johnson, 1987). In multiethnic communities, various linguistic resources manifest to indicate an ethnic identity such as the use of different varieties, CS, and the use of the standard language, among others (Fought, 2006). For example, people may switch codes to negotiate their dual identities as bilinguals and the rights and obligations indexed to linguistic choices within their community of development. This is done to influence the outcome of interactions, as speakers choose the code based on their assessment of the situation (Myers-Scotton, 1995).
Ethnic labels can also predict certain attitudes towards language. In a study conducted in New Mexico with Spanish as a Heritage Language Learners (SHLLs), Henderson and colleagues (2020) noted that those who identified as Mexican-American, Hispanic, or Spanish exhibited higher ratings of the “value” dimension of Spanish, defined as “the public display of language use to establish a social group identity” (p. 29), compared to participants who did not use those labels. The authors indicate:
When individuals no longer identify with a specific culture, there may be less motivation to maintain a linguistic link to that culture […] SHLLs who do not identify with a specific Spanish-speaking culture may have different language attitudes towards Spanish than peers who do identify with a specific culture. (Henderson et al., 2020, p. 29)
Identity negotiations through language are dynamic and context-specific, with shifting stances in monolingual and bilingual interactions, challenging static cultural and social categories (Bailey, 2007). The current study aims to determine whether this connection between culture and linguistic features may be observed in the bilingual community in North Carolina.
2.2. Ethnic Self-Identification
Fought (2006) remarks that in the construction of ethnic identity – both self-identification and the perception and opinions of others – are important. Phinney and colleagues (2001) found the importance of parents, ethnic language (referring to the original language of their migrant parents), and peers’ ethnicity to be factors for the ethnic identity of adolescents.
Aligned with that argument, Herman and Castilla (2010) note that peer groups play a significant role in shaping racial/ethnic adolescent identities. Conversely, the financial and educational conditions of parents impact social interactions, among other factors, which influence young individual identities (Iceland & Wilkes, 2006). As a result, minority families with limited financial resources tend to live and interact with people from similar backgrounds, which leads to the maintenance of the community’s identity.
Regarding the role of ESI, context may shape the label used. Individuals can employ a variety of labels depending on with whom they are communicating, and the language used. Dowling (2005), for example, explored the self-identification of individuals of Mexican descent, investigating the reasons underlying respondents’ preferences regarding the terms Mexican-American, Mexican, Mexicano or Mexicana, and Texan, Tejano or Tejana as their ethnicity or race for identification. Via interviews with 52 speakers, the author reported that the use of labels in English and/or Spanish were not equivalent in translation, each having distinct implications not only in each language (Spanish vs. English), but also in different sociolinguistic contexts. Among the U.S.-born participants, most reported identifying as Mexican-American given that they a) were not born in Mexico, b) sought a way to distinguish themselves from Mexican-born immigrants, and c) desired to express their citizenship and pride in the U.S. The majority of this same group of speakers rejected the term Mexican but accepted Mexicano in certain contexts because the former term refers to someone from the border, while the latter is related to their ancestors. In contrast, among immigrants, as well as individuals born in the U.S. but raised in Mexico, the term Mexicano was primarily used. Even those identifying as Mexican-American tended to identify as Mexicano when speaking Spanish, showing the importance of considering both the language and the context of ESI terms.
As noted, studies have found that G2 individuals navigate a complex web of ethnic and linguistic identities. Geerlings and colleagues (2015) conducted surveys and interviews in English with adolescents from Miami/Ft. Lauderdale and San Diego. The authors noted that adolescents tend to identify themselves with a hyphenated-American or pan-ethnic label more than with their parents’ origin, and they prefer English terms over descriptors in Spanish. This tendency increases over time, while their preferences for their heritage language decrease. In a study with similar findings conducted in Los Angeles, Fuligni and colleagues (2008) observed that G2 individuals have a greater tendency to use pan-ethnic and American terms compared to G1 individuals. Additionally, G2 adolescents who were not born in Mexico felt unable to primarily consider themselves Mexican.
As explained, self-identification may be adapted according to the context. In this regard, the interviewer may exert influence on the message conveyed during an interview (Giles et al., 1991; Labov, 1972). It is possible that interviewees accommodate their responses and linguistic features because the message is inherently tied to the presence of the receiver (Giles et al., 1991). Similarly, identity is expressed according to multiple factors, including the interaction in which it emerges, and may be either deliberate or unconscious (Bucholtz, 2005; Bucholtz & Hall, 2008). Therefore, it is important to acknowledge that researchers play a significant role in interviews, and the fact that both interviewers in the present study were native Spanish speakers must be considered as a possible influence of the results obtained.
In summary, ESI is influenced by both external perspectives and contextual factors, and a trend exists among G2 and G3 bilinguals to distance themselves from the ethnic heritage of their migrant family that extends to language usage as well. The present study explores whether this observed tendency, concerning the evolving dynamics of language usage (i.e., choice of Spanish, English, CS) and ethnic identity, is manifesting in NC.
2.3. Language Contact Phenomena
While the outcomes of language contact are evident in various parts of the bilingual linguistic system, the present study focuses specifically on CS and its connection with identity. Considerable research has established that CS is present in stable bilingual communities and demonstrates the speaker’s competence in both languages as it requires linguistic skills to express oneself (Poplack, 1980). This view has not always predominated, however, and Myers-Scotton (1995) reports that previous scholars argued that CS was a “part of the performance of the imperfect bilingual” (p. 47).
Various proposals exist regarding what does, and does not, constitute CS. Poplack (1980) categorizes CS into intrasentential, intersentential, and tags, and argues that intrasentential CS are the most complex to master due to their structural complexity. Torres and Potowski (2016) also include discourse markers (DMs) such as you know, I mean, so, like, and well, as types of CS, along with intersentential, one-word intrasentential, and two-or-more word intrasentential. As DMs have been reported to constitute a large proportion of one-word CS and, in fact, are “often the most frequent type of lexical borrowing in the Spanish of bilinguals” (Torres, 2011, p. 494), the present analysis will consider DMs as a type of CS, aligning with the classifications of Torres and Potowski (2016).
With specific reference to research on CS in bilingual communities throughout the U.S., several investigations have revealed notable trends regarding the frequency and types of CS among speakers of different sociolinguistic generations. Torres and Potowski (2016) evaluated CS in the Latino community in Chicago comparing different generations and ethnic groups. They found that G3 speakers from Puerto Rico had the lowest proficiency in Spanish, yet they produced the apparently most complex CS (intrasentential switches with two or more words) compared to G2 speakers and other G3 ethnic groups. The authors stated that the motivation for CS is not only proficiency; it is also determined by the ethnolinguistic group to which the individual belongs. For example, Puerto Ricans, notwithstanding their generation, used more CS than Mexicans. Additionally, in the case of the G3 participants, CS occurred because they felt more confident communicating in English. The authors also note that, as proficiency in Spanish decreases, the use of English increases. In this way, what may serve as a community identity marker for some speakers may primarily serve to support lower proficiency for other speakers.
With specific reference to NC, little research has been conducted on CS. To date, only Ronquest and colleagues (2020) and Michnowicz and colleagues (2023) have assessed the use of DMs and CS among G1 and G2 individuals in the state. Regarding DMs, G2 speakers exhibited a greater integration of English DMs into their speech in comparison to G1 immigrants (Ronquest et al., 2020). Second generation speakers were also found to produce the most complex CS compared to G1, although the vast majority of the switches consisted of single words (Michnowicz, Ronquest, et al., 2023).
Additional research in NC points out that while Spanish has not been completely displaced, young people have fewer contexts to use Spanish outside their family settings (Michnowicz, Trawick, et al., 2023). The G2 speakers in NC have primarily shifted to English possibly due to the lack of cyclical bilingualism characteristic of other U.S. regions.
3. Research Questions
One of the many facets of human identity is ethnicity, and both self-identification and the perception of ‘the other’ are relevant in the construction of ethnic identity. Additionally, evidence suggests a relationship between ethnic labels and attitudes towards language. Furthermore, regarding ESI labels, it is suggested that the linguistic context influences the meaning assigned and the label chosen. Within a diverse community, individuals utilize linguistic resources to express their ethnic identity, including the use of standard language or CS. Code-switching appears to have various motivations and levels of prestige, but it is also influenced by social identity, as individuals may seek to express a dual identity. In the present study, ESI will be evaluated alongside CS as a possible indicator of dual identity and contact phenomena within the bilingual community in NC. Based on the aforementioned concepts, the present study will address the following research questions:
-
What ESI labels do NC bilinguals use, what are their meanings, and do these labels change according to different contexts (e.g., if speaking in English, Spanish, or within a predominantly Hispanic or Anglo community)?
-
Which language do interviewees select or prefer in daily life, as well as in the context of the interview?
-
Do NC bilinguals employ CS, and if so, what specific types of CS occur in their speech?
-
What is the relationship between ESI, language choice, and CS? That is, is there a connection between how participants self-identify, language selection, and the types of CS they exhibit in their speech?
4. Methodology
Data regarding ESI, language selection, and CS were obtained via semi-guided sociolinguistic interviews lasting approximately 20 minutes that were conducted in the spring of 2023, both online and in-person.[2] Participants were recruited through the personal and professional contacts of the first author as well as through snowball sampling. Given the method of recruitment, the sample is somewhat homogenous. Participants were given the freedom to respond in Spanish, English, or a mixture, and they were encouraged to use the medium with which they felt most comfortable. Only one interviewee (“Hugo”) requested that the questions be entirely in English, and one other participant (“Fito”) used more English as the interview progressed, even though the questions remained in Spanish.
Four women and four men between the ages of 18 and 22 participated in the interviews. Seven participants were born in North Carolina, while one was born in Illinois. All were residing in NC at the time the interviews were conducted, with seven of them having lived there since birth and one moving to the state at the age of six. Seven of the eight participants indicated that both parents were born in Mexico and were therefore classified as G2. One participant indicated that his mother was born in Los Angeles with Mexican and Salvadoran ancestry, while his father was born in Mexico, and was therefore classified as G3:1. The diversity of the sample is justified given that this is an exploratory study aiming to reflect the heterogeneity of the bilingual population in NC. All interviewees were students, with seven enrolled in university and one in high school. In terms of parents’ education, all reported that they attended school, with some completing it, but none pursued higher education (See Appendix A, Table A1). The interviews were conducted by two native speakers of Spanish (one Mexican male and one Peruvian female, see Appendix A, Table A2).
Both interviewers were present throughout the entire interview, took turns asking questions, and did not engage in CS. The interview questions focused on exploring the choice to speak English and Spanish in different contexts (e.g., familial, social, academic), as well as inquired about participants’ opinions on the use of Spanish in NC. Several minutes into the interview participants were asked “¿Cómo te identificas cultural y étnicamente?” [How do you identify yourself culturally and ethnically?]. All interviews concluded with questions concerning demographic information. Conversations were recorded using a cellphone in the case of in person interviews, or via Zoom in the case of online interviews. One of the interviews was orthographically transcribed manually, and the rest were transcribed using the Speak AI platform (Bryden, 2019) and then corrected for accuracy by the first author.
Participants’ responses were organized thematically by topic and further analyzed to identify patterns in the discourse. The analysis of CS was conducted by evaluating the responses and categorizing them as presented in Table 1.
Compound nouns were coded as two or more words. Proper nouns and place names were not considered CS unless there was a known Spanish name (e.g., Spain vs. España). Each switch counted as one token, meaning that if a participant switched three times during the same utterance, all three were counted as switches.
The relationship between language proficiency, language choice, and CS has been discussed by several authors, however, they indicate that proficiency is not the sole predictor of frequency or complexity of CS (Gross et al., 2022; Ribot & Hoff, 2014; Sinclair & Fernández, 2023; Torres & Potowski, 2016; Tulloch & Hoff, 2023). As the present study obtained a more global measure of proficiency based on responses to interview questions and not a quantifiable proficiency score, we are unable to directly assess the role of Spanish or English proficiency in the present study. We acknowledge that this precludes our ability to systematically address questions specifically pertaining to proficiency but seek to include such measures in future research.
5. Results
5.1 .Ethnic Self-Identification (ESI)
We begin by exploring the ESI labels adopted by NC bilinguals, their meanings, and if these labels change according to different contexts (e.g., if speaking in English, Spanish, or within a predominantly Hispanic or Anglo community). (See Appendix C for detailed information).
In response to the question “¿Cómo te identificas cultural y étnicamente? ¿Y cuando hablas con alguien que no habla español?” [How do you identify yourself culturally and ethnically? And when you speak with someone who doesn’t speak Spanish?], five participants answered Mexicana or Mexicano, with some variation. One answered Mexicana-Americana and another responded with the English label Mexican-American, despite the question having been posed in Spanish. One participant answered Chicano. Regarding the contexts where they use each term, out of the five people who initially mentioned Mexicana or Mexicano, only one stated that they only used it in the U.S.; in Mexico, they use Americano, showing the importance of place and context regarding ESI.
Further conversation encouraged the participants to include other labels and explain their choices. Regarding the meaning of different ethnic labels, and specifically Mexicano, attitudes were overall positive or neutral. Some participants conceptualized this label in relation to their parents and their heritage and traditions. Others, however, mentioned they noticed prejudice from others when using the term Mexicano, but this does not interfere with their usage of this label.
The use of ethnic labels was explored further, inquiring how participants self-identify in different contexts. As an example, “Gracia” (female, 22 years old), when discussing differences identifying herself when interacting with Americans, stated “Diría que soy Mexican-American, yo creo que por tanto tiempo con la clima política también hay una, like, uhm un negative connotation si dice oh, I am Mexican…” [I would say that I’m Mexican-American, I believe that with the political climate for so long, there’s also a, like, um, a negative connotation if you say, Oh, I am Mexican…].
Among those who identified as Mexican, some expressed rejection of the term Mexican-American. “Brenda” (female, 18 years old), for example, explained: “A mí no me gusta cuando me dicen like americana y no me gusta el término de mexicana-americana, no me gusta I mean, yo soy 100 % mexicana aunque soy nacida aquí […] yo siento que este país tiene bastantes oportunidades […] pero nunca han hecho algo que no discrimine de mi familia…” [I don’t like it when people say like American’ to me, and I don’t like the term ‘Mexican-American.’ I mean, I’m 100% Mexican even though I was born here […] I feel like this country has plenty of opportunities […] but they’ve never done anything that doesn’t discriminate against my family…].
“Gracia”, who initially mentioned using Mexicana-Americana also indicated using Mexicana when interacting with “Latinx groups.” Some acknowledged that while the label Mexican-American technically applies to them, they prefer to identify as Mexicana or Mexicano. As an example, “David” (male, 20 years old) explained: “Mmmm la verdad nunca, nunca le he puesto mucho mucha atención porque en sí soy Mexican-American, pero I don’t know, soy mexicano orgullosamente mexicano so.” [Hmm, honestly, I’ve never really paid much attention to it because I’m Mexican-American, but I don’t know, I’m Mexican, proudly Mexican, so].
Participants in the present study generally did not report differences in meaning or usage between English and Spanish ESI terms, although the context continued to play a role. For example, the use of Mexican-American revealed the practical fact of citizenship, not any pride associated with it, and conversely, in some cases, the lack of Mexican citizenship inhibited them from using Mexicano.
As an example, “Fito” (male, 18 years old) stated that “We are Mexicans but since we are from here, I feel like it is not the same as if you were from Mexico […] we were born in the U.S. […] we have roots from Mexico […] I feel like we have to include the American part.”
Another participant, “Gracia” (female, 22 years old) stated “…yo, siempre me he identificado como Mexican-American, porque this is porque soy de padres mexicanos, pero nací aquí, so yo creo que es importante de tener las dos Nationalities en tu identity […] yo nunca he ido a México para decir eso.” [I’ve always identified as Mexican-American because this is because I’m from Mexican parents but was born here, so I think it’s important to have both nationalities in your identity […] I’ve never been to Mexico to say that].
One participant, “Hugo” (male, 20 years old) explained in English the use of Chicano as a cultural identification and elaborated: “…for me I, I would say it’s […] having both and holding both of those near to you, you know, and, and being able to, to, to really be both […] you can just comfortably be both and it’s not like an issue you know…”.
Additionally, some participants used other labels: three participants mentioned Hispanic or Hispana, two reported Latina, and one introduced Latinx. These participants report maintaining a connection to Mexican culture through their family and traditions, listening to Mexican music, eating Mexican food, and, in some cases, visiting Mexico. Regarding gender, all four female participants reported using the term Mexicana, among other labels. The male participants displayed more variability in their responses: two reported Mexicano, one reported Mexican-American, one Chicano.
To summarize, six individuals reported identifying as Mexicano or Mexicana in their interactions. Additionally, two participants indicated using Mexicanoamericano or the equivalent term in English. One person reported using Americano solely while in Mexico, and another participant reported using Chicano. Furthermore, six participants mentioned employing multiple ESI labels, often depending on the context or language being used. Participants did not report differences in meaning or usage between English and Spanish terms, and women tended to identify as Mexicanas more frequently than men.
5.2. Language Choice and CS
In this section, we discuss the participants’ language preferences in daily life and in the interview and the types of CS employed. Out of the eight participants, seven reported using Spanish with their families, primarily with older people. While all participants engage with friends who speak Spanish, they reported speaking Spanish with them only sparingly. Regardless of predominantly using English in their university and professional environments, four participants mentioned instances where they’ve used Spanish, particularly in jobs involving services with Hispanic individuals or during Spanish classes (four had taken Spanish classes).
Despite having limited opportunities to practice Spanish, all participants acknowledged its importance for their professions and life in NC, and reasons given included the ability to assist the community and access better job opportunities. Participants expressed comfort with using a mix of languages during their interactions with family and friends, with their reasons ranging from forgetting a word to incorporating idioms or simply finding it amusing.
Moving on to the choice of language used in the interviews, as well as the types of CS present in participants’ speech, Table 2 interfaces language of choice with CS.
As shown in Table 2, participants can be grouped into those who 1) spoke only Spanish throughout the interview and did not engage in CS; 2) maintained most of the conversation in Spanish with variable CS; and 3) spoke primarily English and produced very few examples of CS. This grouping, along with ESI, will be explored in greater detail in section 5.3.
Focusing on the types of CS present in the interviews, the most frequent were DMs (57 tokens, 38.78%), followed by one-word intrasentential (52 tokens, 35.37%), intrasentential switches consisting of two or more words (34 tokens, 23.13%), and intersentential (full phrase) (4 tokens, 2.72%). Four of the six participants who handled the conversation in Spanish produced 138 of the 147 (93.87%) total CS.
The use of English DMs – the most common type of switch – are exemplified by “Gracia” in her explanation about a job she had in a doctor’s office, and “Cesar,” describing the usefulness of Spanish in his community, respectively:
“Allí, en esa área, hay muchos hispanos. So, la mayoría citaba yo, so, mucho español con ellos.” [There, in that area, there are many Hispanics. So, I often helped them with their appointments, so, I spoke a lot of Spanish with them].
“…porque ahorita ahorita todavía hay muchas personas. You know ya mayores que para ellos no quieren inglés…” [Because right now, there are still many people. You know, older people who don’t want English].
One-word intrasentential switches were also frequently noted in the discourse of several speakers. “Brenda”, for example, when describing the place where she lives, stated:
“…es como un pueblito pequeño está como más para el norte como dando el border con Virginia…” […it’s like a little town, like more to the north on the border with Virginia…].
“Hugo”, when discussing the use of Spanish when watching television, switched into Spanish, to refer to the cultural phenomenon of telenovelas (soap operas):
“… I’m not like shy, shy from watching like Spanish, any anything in like Spanish, you know? It’s not like weird for me. Definitely if they had novelas…”.
Intrasentential switches of more than two words were also present in the data. “Gracia” provided the following switch from Spanish into English, and also included the discourse marker “so” as a transition:
“…mi primer trabajo era un cashier en Food Lion y entonces me promoted me to a supervisor so hay veces…” […my first job was a cashier in Food Lion and then they promoted me to a supervisor so sometimes…].
Finally, two participants displayed intersentential CS, uttering one entire sentence or thought in one language, and then switching completely to another language, as is evident in the following quote from “Fito”:
“Ah. Déjame pensar esta. To be honest, I think that being Mexican is to have, amm to be honest, to have parents …” [Ah. Let me think of this one. To be honest, I think that being Mexican is to have, amm to be honest, to have parents…].
In summary, most participants reported choosing to communicate with their families in Spanish and stated that they have limited opportunities to use the language outside the home. They value Spanish both in terms of how it contributes to their lives as members of the community and appreciate the opportunities it offers them in the future. They interact primarily in English in most contexts and feel comfortable using a mix of languages during their interactions with family and friends. In the context of the interview, six speakers engaged in some kind of CS, the most common type consisting of DMs.
5.3. Language Selection, CS, and ESI
Finally, we explore if there is a connection between participants’ ESI, their language choice, and the specific types of CS present in their discourse. As observed shown in Table 3, there is diversity in the responses regarding the ESI labels used by each participant and sub-types of CS.
Regarding general patterns, two participants (Group 1) maintained the conversation in Spanish and did not engage in any CS; both identified as Mexicana and Mexicano, among other labels. Four participants (Group 2) maintained the conversation in Spanish, with varying degrees of CS and all also identifying as Mexicana and Mexicano, among other labels. Two male participants, neither of whom identified themselves as Mexicans, used English during the interview with minimal CS (Group 3). “Fito” began to use English partway through the interview although he did attempt to maintain the use of Spanish, albeit marked by pauses and hesitations. Despite his preference for English, the interviewers continued to address him in Spanish. He produced a variety of CS, with intersentential (phrasal) switches the most commonly observed in his speech (3 out of 7 total switch types). “Fito” is a G2 individual, the youngest interviewee, and the only one in high school, who reported self-identifying as Mexican-American. “Hugo” requested that questions be posed in English. Although questions and answers were conducted in English, during the first three questions, the participant attempted to respond with short phrases in Spanish and integrated two intrasentential switches. He is a G3:1individual, which may be the fundamental reason for his preference to speak English, and he was the only participant who switched from English into Spanish. Additionally, “Hugo” primarily self-identified as Chicano, claiming allegiance to two cultures rather than identifying as half of both.
In summary, English was used more frequently among the two participants who did not report identifying with the label Mexican or Mexicano/a in their daily interactions. These speakers also employed very few examples of CS. Conversely, those who reported identifying as Mexican at some point in their daily interactions were more inclined to maintain the conversation in Spanish, either entirely or with some instances of CS.
6. Discussion
The first objective of the present study was to describe the ESI labels used by the interviewed individuals. Even though all participants were born in the U.S., the majority indicated using the term Mexicana or Mexicano. Furthermore, their responses regarding ESI labels are diverse and reveal subtle differences depending on the context. This finding aligns with what has been described about the complexity of identity and the multiple factors that may influence each individual (Fought, 2006; Herman & Castilla, 2010; Iceland & Wilkes, 2006; Phinney et al., 2001).
Despite the diversity, however, the most frequently adopted term was Mexicana or Mexicano, reflecting strong ties with their heritage country. The overt connection to their parents’ culture contrasts with studies in other regions on G2+ individuals (Dowling, 2005; Fuligni et al., 2008; Geerlings et al., 2015) and may relate to the formation of a new bicultural community in NC (Michnowicz, Ronquest, et al., 2023; Michnowicz, Trawick, et al., 2023; Ronquest et al., 2020). Dowling (2005), for example, reported that in Texas, the label Mexican was primarily limited to those participants who were Mexican-born. The trend in the interviews among the NC Latinos in the present study is the opposite: they prefer Mexicano despite being born in the U.S. Of additional importance, Dowling’s study revealed a rejection of the term Mexican and the different meanings associated with Mexican and Mexicano, which was not mentioned in any case in NC. The tendency of individuals of migrant descent to claim a bicultural ESI in places such as Texas, Miami/Ft. Lauderdale and San Diego, and Los Angeles (Dowling, 2005; Fuligni et al., 2008; Geerlings et al., 2015, respectively), and the varied meanings of the ESI labels according to the language choice, contrast with findings of the present study, where the tendency was to self-identify as Mexicana or Mexicano. Thus, their connection and association with the ethnic labels differ from those residing in more established dialect regions who prefer the term Mexican-American. That is, while research in more well-established bilingual communities has identified clearer patterns or preferences regarding identity and language use, the diversity reported among the participants included in the present study likely reflects an earlier stage of development in which bilinguals are still negotiating their ethnic, cultural, and linguistic identities (Michnowicz, Trawick, et al., 2023).
However, we acknowledge that while these differences may reflect the divergences with those communities (i.e., a longer history of migration, a border with Mexico), they may also be due to the size and nature of the sample. Dowling (2005), for example, evaluated different generations and ages, whereas Fuligni and colleagues (2008) interviewed and utilized a survey instrument in English. Future investigations regarding ESI, language choice, and CS in the Southeastern U.S. should aim to include larger and more varied samples (e.g., balanced number of G2 and G3 speakers) in order to provide a more nuanced characterization.
In some cases, linguistic or social contexts appear to influence ESI of the sample evaluated, similar to Dowling (2005) and Potowski (2016), where the cultural and linguistic experiences of bi/multicultural individuals serve to negotiate their fluid and dynamic ethnic identity in different situations. In this regard, a factor to consider in the interpretation of the present results is that the language (Spanish) and the origin of the interviewers may also have influenced the ESI labels reported by the interviewees.
Moving on to the discussion of the choice between speaking English, Spanish, or CS, all interviewees reported that English is the language they speak in academic and work settings, with little opportunity to use Spanish outside of family contexts, similar to the findings reported by Michnowicz and colleagues (2023). Participants also described that they are comfortable mixing English and Spanish languages during their daily lives when speaking. Participants described that they value Spanish as a bridge to their community, for their future, and for North Carolina. In addition, even those who did not identify themselves as Mexican, or who used more English in the conversation, expressed a connection with their heritage culture. Our findings align with those of Henderson and colleagues (2020), who reported that individuals who rated Spanish highly were also those who self-identified with their parents’ culture.
In terms of CS in the interviews, six participants switched between languages, and the distribution of switches observed (DMs > one-word intrasentential > two-word intrasentential > intersentential/phrasal) is consistent with Torres and Potowski (2016) for G2 Mexicans. It is important to note, however, that Torres and Potowski described that Chicago speakers’ use of CS varied considerably across different sociolinguistic generations and ethnolinguistic groups. While the present study does allow for some insight into how language choice and CS among bilinguals from NC compares with other regions, the size of our participant pool does not allow us to make definitive conclusions. Additionally, our study included seven G2 and one G3:1 participant, which we decided to include in the analysis to reflect the diversity of the bilingual population in NC. On a related note, Michnowicz and colleagues (2023) in NC found that single word switches were prevalent in CS among G1 and G2 speakers, but also noted a higher proportion of intersentential switches than the present study. These differences may stem from the analysis of an additional ethnolinguistic group (i.e., Central Americans in addition to Mexicans), the Spanish proficiency of the participants, or due to the influence of the interviewer: in the present study, interviewers were native speakers of Spanish, however in Michnowicz and colleagues’ study, interviews were directed by mostly L2 speakers of Spanish. Despite the fact that participants in the present study were encouraged to speak their preferred language throughout the interview, the fact that the interviewers spoke primarily in Spanish and did not engage in any CS themselves may have influenced some participants to stay in monolingual mode, effectively reducing or eliminating CS during the interview.
Regarding the final objective of our study, we wanted to explore the possibility that CS may serve as a means to express dual identities (Fought, 2006; Myers-Scotton, 1995). Given that six of the eight participants engaged in some form of CS during the interviews, there appears to be an overall trend in the data, however, there is no clear pattern between the ESI labels used and the specific sub-types of CS. Two participants did not produce any CS at all (identifying as Mexicanos), while two others engaged in very few instances of CS (identifying as Mexican-American or Chicano). However, identities are not always explicitly claimed and may sometimes be expressed implicitly through speech (Bailey, 2010). Not all the interviewees who codeswitched explicitly identified as Mexican-American, and not all the interviewees who identified as Mexican refrain from CS. Even though all the interviewees who maintained the conversation in Spanish identified as Mexicana or Mexicano, those who reported not using these labels tended to use more English. At present we cannot completely determine if CS is a practice utilized as a means to overtly express or reinforce a dual identity or if it is solely used by some speakers due to lower proficiency levels in either English or Spanish. We save more in-depth analyses regarding this topic for future research.
In summary, the interviews with our participants revealed the diversity and complexity of the newly forming bilingual community in the Southeast, highlighting the intricate connections between ESI, language choice in specific contexts, and patterns of code-switching. The general patterns and findings align with the understanding that language usage plays a crucial role in shaping ethnic identity (Giles & Johnson, 1987). Ethnic labels can predict attitudes toward language use or choice, and when individuals identify with a specific culture, there may be more motivation to maintain a linguistic link to that culture (Henderson et al., 2020).
7. Conclusions
The implications of this study lie in its contribution to the ongoing research aimed toward understanding the evolving bilingual communities in the Southeast. Expanding our knowledge of identity construction and language usage can offer valuable insights into heritage language instruction, ultimately improving Spanish courses for heritage speakers similar to those who participated in this study. In more general terms, studies of identity and bilingualism carry broader implications and raise awareness about bilingual populations and communities. As Zentella (2008) states these aspects “have implications for how multilinguals are seen and treated” (p. 3), and studies on this topic may result in more appropriate and culturally sensitive approaches to addressing bilingualism and language mixing among heritage populations. Future studies involving a greater number of participants from multiple sociolinguistic generations and ethnic backgrounds, as well as those that incorporate other potential factors affecting the language choice such as proficiency and language attitudes, will grow our knowledge of the complex nature of the relationship between language and identity.
Acknowledgements
We would like to thank three anonymous reviewers for their helpful feedback and comments on this work. We also thank Carlos Dominguez for his assistance in this project.