Sociolinguistic Style, Awareness, and Agency among Southern California Latinx Spanish–English Bilinguals

Abstract

This study examined different degrees of awareness regarding the stigmatization of Southern California (SoCal) Spanish across four groups of Spanish–English bilinguals from Southern California (n = 87). The participants were presented with Spanish sentences and asked to decide which profile of speaker would likely express that sentence, given six options, such as: “someone living in Los Angeles/SoCal who grew up in Mexico” or “a Spanish-English bilingual who grew up in Los Angeles/SoCal”. Experimental stimuli included seven different linguistic categories of stigmatization, including English contact forms. The participants tended to attribute the stigmatized forms to bilinguals who grew up in Southern California. Central Colloquial and Taboo categories were more salient and perceived as forms used by people in Mexico. In contrast, English borrowings and redundancies seemed to be recognized by the participants, particularly for simultaneous bilinguals who grew up in Southern California, as salient forms of an identified Southern California Spanish variety. The results are interpreted within Exemplar Theory, with certain stigmatized forms indexing “Mexican Spanish” exemplars, and English borrowings identified as exemplars of SoCal Spanish. We advocate for usage-based approaches to understanding language perceptions and critical approaches to interrogating academic discourses.

1. Introduction

Historically, U.S. Spanish has been confined to a linguistic borderland and regarded as “deficient” and marginalized among Spanish varieties (e.g., Anzaldúa 1987; Christoffersen 2019). Speakers from Spanish-speaking populations in Latin America and Spain have long maintained negative attitudes toward the effects of contact with English, including adapted and non-adapted lexical borrowings (Moreno de Alba 2003). Some scholars have contested the presupposition of Spanish in the U.S. as a “simplified”, “unstable” mixed variety, among other sociolinguistic ideologies, and have demonstrated its complexity and contextualized systematicity (Bessett and Carvalho 2021; Calafate 2022). Nevertheless, these reductionist and marginalizing biases continue to permeate and perpetuate research discourses, particularly in the U.S., with problematic pedagogical implications.

Since colonial times, the Spanish spoken in Southern California (SoCal) (and its speakers) has been marginalized not only because of the English elements it often incorporates, but also because it includes many forms from Mexican and Central American Spanish that are considered “informal” and/or “non-standard”. This variety has also been perceived as nonexistent due to language ideologies, including raciolinguistic ideologies co-naturalizing race and linguistic practices (Flores and Rosa 2015), as well as a lack of knowledge and proper consideration of historical sources (Lamar Prieto 2014). California Spanish, similar to other border and Southwest Spanish varieties, has been characterized under traditional dialectological studies as a “rural” and “illegitimate” variety incorporating historical linguistic elements diverging from the “educated norm” (Lamar Prieto 2014; see also Moreno de Alba and Perissinotto 1998). These kinds of previous dialectological studies in Spanish in the Americas have been criticized because of their limited focus on national varieties within minimalist categories that do not differentiate between regional and more generalized elements and because they often center on descriptions of phonology and lexical forms (Lipski 2011).

More recent sociolinguistic and variationist studies focusing on the contact of Spanish with English in the United States have challenged these traditionally understood geographical isoglosses and have focused on social traits such as ethnic identity formation, discourse, and pragmatics (e.g., Aaron 2004; Galindo 1999; Mendoza-Denton 2008). Other studies have underscored the lack of analysis of conflict and social differentiation within the context of colonization in the Americas (see Rosenblat 1990). For instance, Parodi and Luján (2014) suggest that Spanish varieties on the American continent developed from intense contact and conflict with Indigenous languages and cultures. After the genocide of many Indigenous people through violent means and the introduction of European diseases to the continent and the seizure of much of the territory in the Americas, Spaniards still found themselves in the minority (Parodi and Luján 2014). This is noteworthy since, to explain the formation of new cultural and sociolinguistic realms in the Americas, Parodi and Luján (2014) observe that the new Spanish varieties constitute a part of cultural chronotopes, that is, mental spaces comprising individuals’ and communities’ self-consciousness as “the other” within which their identities reside. The concept of cultural chronotropes derives from the original definition of chronotrope by Bakhtin (1982) as a unit of both space and time, in which a new and differentiated community, including criollos—those directly descending from Spaniards—developed dialogic relationships (Parodi and Luján 2014; see also Bakhtin 1982). In this dialogic relationship, two cultures and experiences can exist simultaneously and in the same virtual space, differently from real time and space. Thus, Parodi and Luján (2014) propose the concept of the new mestizo cultural chronotope to study the Spanish language on the American continent, in which Spanish-speaking communities coexist in dialogue within their own kind (e.g., Spaniard colonizers with other Spaniard colonizers) but, more importantly, in contact and conflict with the “other” (e.g., Indigenous people adopting Spanish lexical elements). In this way, self-consciousness regarding their own alterity was fundamental to the sense of belonging and sociocultural group membership of people in the Americas. Nevertheless, this analysis missed the opportunity to further elucidate the racial and classist ideological differentiation processes and the internalization of these processes into the sociolinguistic mestizo cultural chronotope configuration.

The fact that the analysis of Spanish sociolinguistic ideological configurations and speakers’ recognition of their own sociocultural and linguistic varieties has been peripheral in academia is not surprising (see Del Valle 2014). Dominant and influential precursory studies have been strengthened through nationalistic institutions and by individuals themselves. As Agha (2006) exposes,

[u]nder conditions of nationalism and language standardization locale-specific speech varieties can be held up as normative models for speakers in other locales so that language users (and linguists) whose institutions are trained by standard setting institutions can come to find themselves inclining to the view that normatively defective speech does not in fact occur (or is ‘ungrammatical’ or whatever); when successful, institutionalized efforts of these kinds may even bring more and more of the discursive practices of language community into conformity with a single model, thus effectively erasing the divergence between model and reality, at least for a while. These processes are themselves interesting features of social life that deserve systematic study—both in terms of how they unfold and in terms of the consequences they have, when effective.

(Agha 2006, p. 293)

Moreover, current interrogations of power and marginalization in linguistics have been left to the margins or wholly disregarded as non-scientific “under the guise of wanting to create an objective science of language” (Charity Hudley and Flores 2022). In the case of studies in Hispanic linguistics, most research has been historically conducted at and by the Spanish-speaking cultural hegemonic centers: first Spain and later the metropoles of the former colonies in the Americas. As Martínez and Train (2020) expose, “the fitting of language invention within institutions shapes language-professional practice in numerous ways. It establishes a scientific discourse about language that reflects a particular kind of ‘fluency’ in which empirically observed language practices are made to fit the objectified and artefactualized contours of language invention discourse” (p. 93, see also Martínez 2006). These studies continue to underlie most research designs, contributing to the marginalization of Spanish in the U.S., as these ideologies of language purity and correctness and that racialize and marginalize Latinx bilinguals are not being questioned enough (Higby et al. 2023; Martínez 2006; Silva-Corvalán 2014; Zentella 1998; see also Valdés and Parra 2018).

2. The Imaginary of the Borderlands and Southern California Spanish

In this paper, we analyze how ideological processes are reproduced and continue to evolve and shape Spanish speakers’ perceptions of sociolinguistic practices in SoCal and compare them to other perceived hegemonic cultural centers. Our analysis comes from a position of interdisciplinary critical language complexity (Mufwene et al. 2017), and in which we consider decolonial perspectives in both social and linguistic theory. In this way, we challenge the study of SoCal sociocultural and linguistic practices privileging the apparent dichotomy or pluricentrality of the Spanish-speaking world (see Amorós-Negre et al. 2021), and the U.S.-Mexico borderlands as a periphery (see Duarte-Herrera 2001; Bustamante 1996). We diverge from the problematic framing of the SoCal region as a feral frontier where people do not speak any language or variety “right” by conceiving it as a complex hyperreality for people who inhabit it. And by hyperreality, we do not mean to undermine the material reality of the borderlands but to underscore the necessity to go beyond simplistic, fixed, and essentializing conceptions of speakers and their own social dimensions. Thus, we consider Duarte-Herrera’s reformulation of hyperreality, from Baudrillard (1994), of the U.S.-Mexico border as

the ways in which border representations are used and consumed by different social groups to construct the ‘basic reality’ in which they live and how it affects the ‘basic reality’ of other groups… by exposing the representations of the border as social constructions supported by specific actors that benefit from them, and by exposing the alternative representations of the oppressed, we can call on oblique powers to transform the situation. The relationships of the different levels of representation to the basic reality should be exposed in terms of who bears them and benefits from them in order to de-naturalize them.

(Duarte-Herrera 2001, p. 156)

Additionally, we agree with Duarte-Herrera (2001) that the U.S.–Mexico border definition of “‘[a]n imaginary but formal geo-political divide that regulates the flow of people and products, and that is recognized and enforced by both the Mexican and US states,’ is incomplete and fallacious” (Duarte-Herrera 2001, p. 153) as the definition of isogloss as a two-dimensional representation of a linguistic border is (Barbato 2022). Thus, the problem resides within the Western dominant view in which researchers’ positionalities are immersed. Our study aims to de-naturalize how we understand the borderlands as material and cultural chronotopes because the very actors in Spanish language studies benefit from keeping U.S. Spanish at the margins or erasing it completely.

In this way, we agree that interactional perspectives that focus on the dynamics of language ideologies, attitudes, and perceptions can better explain current mobile communities and their dynamic, fluid, and recursive identities and sociocultural practices (Amorós-Negre et al. 2021; Canagarajah 2012). Our study is centered on a paradigm shift that moves away from the standard language culture and the concept of pluricentrality and focuses on speakers’ language practices, ideologies, perceptions, and attitudes (see Amorós-Negre et al. 2021). Our research aims to analyze further how Spanish speakers in SoCal themselves function as centers as they perpetuate sociolinguistic ideologies at the same time as they exercise their agency and negotiate social meanings, indexing fluid ethnolinguistic identities derived from Greater Mexico and the borderlands’ experiences and social realms.

3. Exemplars in Perception, Awareness, and Salience

The semiotic processes of ideological representations of sociolinguistic difference involve the iconization of speakers and their erasure, and their invisibilization (Irvine and Gal 2000). These perceived deficiencies of racialized Latinx bilingual speakers originate in raciolinguistic ideologies in which language and race mutually conform as a perceived social reality (Flores and Rosa 2015). In fact, the racialization of linguistic practices is an ongoing process that involves both those who produce distinctive linguistic practices and those who perceive them from a particular point of view through time and space and within the structures of power (Chun and Lo 2015). Thus, in the case of speakers of Spanish in the “borderlands”, their iconization and invisibilization occur through non-linear complex processes in which speakers create their own hyperreality with their own sociopragmatic linguistic system. Additionally, speakers may also recognize their iconization and erasure and participate in the ideological processes that reduce their complex practices to unidimensional characterizations. By partaking in this process, U.S. Spanish speakers also project these ideological constructions as intragroup oppositions, a process that Irvine and Gal (2000) call fractal recursivity (see also Holguín Mendoza et al. in press).

These semiotic processes involve sociolinguistic awareness and salience. Awareness depends on salience, which is the degree to which an element “stands out relative to other neighboring items”. The more salient an element becomes, the more speakers will pay attention to it and become aware of it (Drager and Kirtley 2016). Exemplar theory aids in understanding how awareness, salience, and stereotypes of sociolinguistic variables are assembled as cognitive models in which “experiences are encoded in the mind as episodic memories, known as exemplars” (Drager and Kirtley 2016). These are stored cognitively as Exemplar Clouds along with their stylistic social meanings and contexts (Rickford 2016). In this study, we want to understand speakers’ awareness (or lack thereof) regarding several emblematic forms from the so-called “Mexican variety” and if they recognize them as part of their own SoCal variety, contributing to their ethnic identity formation(s).

We follow an exemplar theoretical standpoint in which perception and awareness constitute two distinguishable cognitive processes that involve “noticing” a difference (which leads to awareness) and perceiving a difference (which in the absence of noticing, does not). As observed by Drager and Kirtley (2016), “[p]erception without awareness is possible because many cognitive processes are automatic or reflexive, which contrast with processes that are controlled or reflective” (Drager and Kirtley 2016, p. 1; see also Lieberman 2003). Additionally, awareness of variation is not a requirement for social meaning to be associated with that particular variation. For Choksi and Meek (2016), salience as a type of indexical process establishes a relationship between perception and production. A phrase raises a perception of a particular performance through certain stylistic elements under a sociocultural context, which makes the phrase socially meaningful (Choksi and Meek 2016). That is, “[s]alience emerges through the discursive mediation of performance (production) and the habitual attenuation of interpretation (comprehension/perception) from exposure or experience in everyday life” (Choksi and Meek 2016, p. 249). Salience as a process is more complex and speakers’ experience and context can obscure some interpretations. In a study investigating individuals’ change in their perception of variants, Drager (2011) found that experience has an influence on individuals’ interpretation. Older participants were able to perceive differences between speech stimuli from older and younger speakers, while younger participants were not. Thus, “[w]hile overt awareness of differences may be minimal, if at all, the paths through which relevant activation takes place may be more routinized and established in long-term memory of older listeners than for younger listeners” (Choksi and Meek 2016, p. 250; see also Drager and Kirtley 2016).

In the following section, we discuss a usage-based notion of language and how this perspective supports a critical sociolinguistic analysis of Spanish in the U.S. Research following exemplar theory is compatible with usage-based approaches as they quantify speakers’ experience, allowing researchers to delve into existing sociolinguistic patterns in particular communities (see McGowan 2016). This is not only relevant for our study but crucial since what is on the line is the perpetuation of research methods that continue to iconize and invisibilize entire communities perceived as “homogeneously deficient” and the continuation of a harmful system of language education based on this type of reductionist and limited research analysis.

4. Usage-Based Approaches and Spanish and English in Contact on the Borderlands

In the current study, we employed a usage-based framework to explore speakers’ perceptions of English lexical borrowings in Spanish. Usage-based approaches are broadly defined as models converging notions of language within a larger “functional/cognitive framework” rather than constituting a single unified model (Von Mengden and Coussé 2014). The term borrowing is a linguistic phenomenon that describes “the use of a structure (i.e., phone, phoneme, morpheme, semantic value, or form-function alignment) within a particular linguistic system although it is normally associated with another linguistic system. The term ‘borrowing’ is based on the underlying formal or structuralist assumption that languages are self-contained systems, and that the use of linguistic structures can and should be described in terms of their system affiliation” (Matras and Adamou 2020, p. 237). Within a usage-based framework, borrowings represent the integration of linguistic elements from both languages into a general linguistic repertoire (Matras and Adamou 2020; see also Grosjean 1989). This aligns with research in cognitive psychology demonstrating that bilinguals do not ‘switch off’ individual language systems (e.g., Thierry and Wu 2007; Kroll et al. 2008). Unlike code-switching, a synchronic phenomenon in which speakers use elements from different languages in the same discourse, we adopted a usage-based definition of borrowing as a diachronic phenomenon, in which they become entrenched in a speakers’ or language community’s system through multiple uses over time (Backus 2020).

Within a usage-based perspective, we examine the similarities and differences combining both form and meaning, with a high degree of semantic specificity on lexical items and expressions with pragmatic salience (Backus 2001). In this way, even though it is sometimes difficult to distinguish “new” from “established” lexical borrowings (Backus and Dorleijn 2009), we differentiate between borrowings that have been adapted morphologically and/or phonologically to Spanish from those that retain the source language morphology and/or phonology. The adapted borrowings used in this study have long been registered (Galvan and Teschner 1989), particularly in the Los Angeles/Southern California region. Parodi (2009) defines Los Angeles Vernacular Spanish (LAVS) as a variety descending from Northern Mexico, used mainly by working-class people. LAVS is a Mexican-based koine in a diglossic situation with particular characteristics that differentiate it from other U.S. Southwest varieties, such as those from Chicago, El Paso, or regions in New Mexico (Villarreal 2014; see also Silva-Corvalán 2014). However, as with other Spanish varieties in the U.S., English-influenced elements have been documented in LAVS since early descriptive studies (Phillips 1967). Examples of permanent, adapted borrowings that are generalized include bil (bill) and aseguranza (insurance) (Parodi 2011). Non-adapted borrowings have been documented as innovative sociopragmatic markers, including discourse markers and lexical items functioning as interpersonal markers expressing epistemic stance (Mendoza-Denton 2008; see also Maschler 1998). Under a usage-based perspective, these non-adapted borrowings are understood as part of a continuum in which speakers employ them as strategies of stance and personal relationship mediators (Scheibman 2002; Holguín Mendoza 2018). However, because of these functions, their social meanings are salient/marked and may be perceived as such by speakers (Matras 2009). Non-adapted borrowings may be more salient than adapted borrowings because they do not conform (as much) to Spanish phonological and morphological conventions. We argue that the use of non-adapted borrowings may carry more sociolinguistic meaning than adapted borrowings. Thus, we wanted to explore whether the different groups of bilinguals have awareness of the differences between these two forms, especially whether they are differentially associated with different types of speakers.

5. Sociolinguistic Stylistic Practices “from” the Borderlands and Greater Mexico

In a usage-based approach, linguistic structures are “considered not to be a priori but rather emergent, the product of language use and our cognitive response to our experience with it” (Díaz-Campos and Balasch 2023, pp. 1–2). In this perception study, we aimed to contribute to the understanding of the sociocultural linguistic realm of the bilingual Latinx population in SoCal as a dynamic complex community. Cognitive usage-based approaches have also centered on the semantics of linguistic expressions incorporating morphosyntactic and lexical items within specific contexts and the social meanings and concepts which they encode (see Lemmens 2017). In this study, we framed constructions popularly known as “jargon” or “slang” within a usage-based approach in which “all elements of grammar, be they morphemes, lexical items, idiomatic phrases or larger syntactic structures are considered as symbolic (or constructions) stored in the grammar (hence, an inventory) as long as they are sufficiently conventionalized” (Lemmens 2017, p. 108). Thus, phrases containing rich semantic components, such as those including words or expressions associated with Mexican Spanish, and therefore conventionalized as part of the linguistic practices of the so-called “Mexican Spanish variety” would be recognized by inhabitants of this transnational social realm.

For this study, we also included several forms that have been previously called “popular” or “rural” language spoken within Mexican Spanish and the borderlands (e.g., Aguilar Melantzón et al. 1985; Alarcón 1978; Bills 1975; Espinosa 1915, 1930; Hidalgo 1987; Sánchez 1994). Among these categories, we considered how normative dialectology in the past has merged marginalized linguistic forms in a continuum, including labels such as “criminal argot” (e.g., Coltharp 1965), “Caló jargon” (a variety of U.S.-Mexico border subcultures inherited from the Romani people in Spain, see e.g., García 2005; Webb 1976), and “non-standard/informal/popular” (e.g., Hidalgo 1987; Sánchez 1994) as “Mexicanismos” (e.g., Moreno de Alba 2003) or “Mexican border regionalisms” (e.g., Aguilar Melantzón et al. 1985). A usage-based approach rejects the characterization of forms such as “vato/bato” (from Romani varieties, dude/man), “estar chiveado” (to be embarrassed), “güero” (blond), “dar hueva” (be lazy), “muncho” (variant of mucho, much), and “polecía” (variant of policía, police) as morphological and phonological variants from “popular” or “rural” varieties or as “archaic informal expressions” derived from the “lower classes”.

Thus, an analysis of previously understudied and marginalized forms such as those listed above becomes more revealing if we consider these forms as conventionalized symbolic structures indexing different context-specific social meanings that could be used by any member of the community, regardless of their membership in the socioeconomic (class) continuum or geographic location on a particular side of a national or linguistic “border”, for instance (see Agha 2006). As some studies have already demonstrated, attention to the stylistic dimensions of symbolic constructions reveals the contextualized ideological social meaning creation that leads to people’s metonymic reduction (Agha 2006) in language varieties as emblematic of nations, regions, or other categories of social associations, that is, as an element having the properties of a whole through folk-ideology (Agha 2006). In this manner, particular elements of “Mexican Spanish” (including elements conventionalized as “criminal argot”, “Mexican slang”, “border Caló”, etc.), can be brought to discourse to function as emblems of social difference as part of performances through intentional strategies by the members of social groups to index “Mexican upper-class prestige”, “rural or working-class Mexicaness”, “a Border transnational experience”, “Mexican-American”, or “Chicanx” membership, among the innumerable possibilities for intersectional identity formations (see Guerrero 2018; Holguín Mendoza 2018; Mendoza-Denton 2008).

6. Bilingual Speakers’ Perceptions in SoCal

The present data are part of a larger study of language perceptions and attitudes by SoCal Spanish–English bilingual speakers. In a previous analysis (Holguín Mendoza et al. in press), we found different levels of awareness of stigmatized varieties depending on the age at which bilinguals came to inhabit SoCal’s linguistic milieu. The speakers in the four groups (simultaneous, early sequential, 1.5 generation, and late sequential bilinguals), but especially late bilinguals, showed high awareness of the social meanings of linguistic forms across categories of stigmatization, particularly those indexing “lack of high culture” and “progress”. Thus, SoCal bilinguals understand the social meanings of stigmatization indexed by these language forms. In contrast to late bilinguals, simultaneous bilinguals showed higher awareness of English borrowings, strongly reporting that they had been told that these contact forms were “incorrect”. Additionally, they showed less awareness of stigmatization across the rest of the language categories, such as colloquial Mexican forms, border Caló, patrimonial forms, and taboo expressions. In other words, these words and expressions were more acceptable for simultaneous bilinguals. In this way, simultaneous bilinguals seem to have incorporated these forms into other categories that may stylistically index pride in their Mexicanness, for instance (see Guerrero 2018), in their own sociopragmatic system (Holguín Mendoza et al. in press).

The current study explored to what degree Spanish–English bilinguals in SoCal recognize these same stigmatized forms as belonging to their own community’s linguistic practices or to other Spanish-speaking communities in Spain and the Americas. We also explored how the age of arrival in the U.S. and age of English acquisition modulates this recognition. More specifically, we posed the following research questions:

• Do Spanish–English bilinguals living in Southern California recognize seven different linguistic categories that are classically considered informal or stigmatized as part of their own sociolinguistic practices?
• Do Spanish–English bilinguals living in Southern California differ in their perception of these linguistic categories depending on the age at which they moved to the region?

7. Materials and Methods

Spanish–English bilinguals living in Southern California were recruited and invited to take an online survey administered through Qualtrics. The respondents made four different judgments about a set of 70 Spanish sentences. Each of these judgments was presented in a separate section of the survey. The data presented here are from the first section only, in which the participants were asked who they thought might say the given sentence. The following response options were provided in a fixed order: (1) esta persona vive aquí en el área de Los Ángeles/Southern California pero creció en México (this person is from the Los Angeles area/Southern California but grew up in Mexico), (2) Esta persona es bilingüe en inglés y español y creció en Los Ángeles/Southern California (this person is bilingual in English and Spanish and grew up in the Los Angeles area/Southern California), (3) Esta persona es bilingüe en inglés y español y creció en otra parte de los Estados Unidos (this person is bilingual in English and Spanish and grew up in another part of the U.S.), (4) Esta persona es de México y vive ahí (this person is from Mexico and lives there), (5) Esta persona es de España y vive ahí (this person is from Spain and lives there), and (6) Esta persona es de otro país de habla hispana y vive ahí (this person is from another Spanish-speaking country and lives there). The respondents were allowed to select more than one response choice for each sentence. We decided to include these six options based on the documented linguistic attitudes in Spanish-speaking regions reflecting a hierarchy of superiority: people from Spain are often perceived as speaking “better” Spanish than people in Mexico, and people in Mexico are thought to speak “better” Spanish than speakers in the U.S. (e.g., Guerrero 2018; Hidalgo 1987; Zentella 2017).

The sentence stimuli represented seven different sociolinguistic categories that we delimited based on previous research: redundancies (e.g., bájate para abajo (get down below)), patrimonial/rural forms (e.g., muncho (much)), Mexican Caló (e.g., vato (dude)), Central Mexican colloquial (e.g., güero (blond)), Mexican informal taboo, such as curse words (e.g., güey (dude, interjection)), adapted English borrowings (e.g., troca (truck)), and non-adapted English borrowings (e.g., creepy). Each condition included ten experimental sentences and ten control sentences, which were created by replacing the stigmatized word or phrase with a non-stigmatized word or phrase with a similar meaning. Moreover, we gathered feedback on the experimental forms of the survey from four bilingual speakers living in SoCal to ensure that the phrases were as close as possible to the variety of Spanish spoken in SoCal. In order to assess the participants’ reactions to the target words/phrases, we compared the responses on the experimental items for each stimulus category to the responses on the control items, for which all the other lexical content was the same. To avoid repetition, two stimulus lists were created, with each list containing five experimental sentences and five control sentences per condition for a total of 70 sentences. The lists were counterbalanced across the participants so that each participant only saw one of the two versions of the sentence.

Potential participants were screened for eligibility, which included stated fluency in Spanish and English and age between 18 and 52 years old. We received 91 completed surveys. The data from five responses had to be excluded due to participant ineligibility which was discovered after screening, resulting in 86 responses for the data analysis. Detailed information about the respondents’ language history was obtained prior to the start of the survey. The respondents were divided into four groups based on their language-learning history: simultaneous bilinguals (Group 1, n = 23, mean age = 24.17, SD = 4.61) were born in the U.S. or moved there by age 3 and began learning English by age 3, early sequential bilinguals (Group 2, n = 24, mean age = 25.25, SD = 6.18) were born in the U.S. or moved there by age 7 and began learning English between ages 4 and 7, 1.5-generation bilinguals (Group 3, n = 18, mean age = 27.72, SD = 8.84) moved to the U.S. and began learning English between ages 8 and 12, and late bilinguals (Group 4, n = 21, mean age = 35.24, SD = 10.16) moved to the U.S. and began learning English at age 13 or older. Upon the completion of the study, the participants were provided a small monetary incentive in the form of an electronic gift card.

8. Results

Within each category, we calculated a difference score for the two sentence types (experimental vs. control). Positive values indicate higher response rates for experimental items vs. control items. Negative values indicate lower response rates for experimental items vs. control items. Thus, positive values are interpreted as positive associations between the stigmatized forms and a given response option (e.g., participants selected Esta persona es bilingüe en inglés y español y creció en Los Ângeles/Southern California [This person is a Spanish–English bilingual who grew up in Los Angeles/Southern California] more often with sentences containing adapted borrowings compared to the equivalent control sentences, indicating that they associate adapted borrowings with this speaker group) whereas negative values are interpreted as negative associations between the stigmatized forms and a given response option (e.g., participants selected Esta persona es de México y vive ahí [This person grew up in Mexico and lives there] less often with sentences containing adapted borrowings compared to the equivalent control sentences, indicating that they do not associate adapted borrowings with this speaker group).

Due to the nature of the multiple response categorical variable, the responses were not independent of each other, and therefore inferential statistics could not be used to analyze the data. Instead, we provide descriptive statistics and note tendencies instead of statistically significant findings. First, we examined which stimulus categories were most often attributed to the six different speaker communities provided to the participants. Next, we examined each stimulus category and compared which groups were more and less likely to be identified by the participants as those who would use these forms. Thus, we examined the interactions between groups, stimulus types, and stimulus conditions.

The control sentences showed minimal variation across categories, indicating they served well as a baseline for the experimental sentences. Most control sentences were attributed to people living in Mexico (52–62% of stimuli), followed by people living in Spain and other Spanish-speaking countries (29–37%), followed by people who grew up in Mexico but live in Southern California (21–28%). We looked at the interaction between groups, stimulus types, and stimulus conditions (see Figure 1, Figure 2, Figure 3, Figure 4, Figure 5, Figure 6 and Figure 7). The participants were least likely to assign the control sentences to bilinguals who grew up in Southern California (14–24%) or bilinguals who grew up in another part of the U.S. (9–14%).

When all the stimulus types were combined, the patterns of responses across the four groups of participants were generally similar. The simultaneous bilinguals (Group 1) showed roughly equivalent rates of responses attributing control sentences to speakers living in Spain, to speakers living in another Spanish-speaking country, and to bilinguals living in Southern California who were born there or moved from Mexico (33–39%). The other three groups showed higher rates of assignment for the control sentences to people living in Spain (21–43%) and another Spanish-speaking country (27–31%) compared to bilinguals in Southern California (5–22%).

For all the experimental stimuli combined, all the groups showed lower response rates attributed to people living in Mexico, though the difference between the control and experimental sentences decreased across the groups from Group 1 to Group 4 (Group 1: 14.17%; Group 2: 8.69%; Group 3: 6.03%; Group 4: 5.85%). All four groups also showed lower rates of assignment across all the stimuli to speakers in other Spanish-speaking countries, and this showed a similar decrease in the difference between the control and experimental sentences across the groups (Group 1: 13.04%; Group 2: 7.86%; Group 3: 8.42%; Group 4: 4.90%). Interestingly, Groups 1–3 showed lower rates of attribution for all the experimental stimuli to people living in Spain (26–33% experimental vs. 29–43% control) while Group 4 (late bilinguals) showed higher rates (28.44% experimental vs. 20.82% control). Only simultaneous bilinguals showed lower rates of assignment to bilinguals in Southern California who grew up in Mexico (29.94% experimental vs. 38.76% control). All four groups showed a higher attribution to bilinguals who grew up in Southern California.

8.1. Adapted Borrowings

For the experimental sentences, adapted borrowings were mostly assigned to bilinguals living in Southern California (39.53% experimental vs. 16.74% control) and bilinguals living in another part of the U.S. (16.05% experimental vs. 9.30% control) (as indicated by the positive values in Figure 1, which represent the difference between the response rates for the experimental and control sentences). These forms were not assigned often to people living in Mexico (33.02% experimental vs. 60.93% control). Even though all the groups tended to attribute adapted borrowings to bilinguals who grew up in Southern California, the size of this tendency decreased across the groups, with the simultaneous bilinguals showing the largest preference (32.17% difference) and the late bilinguals showing the smallest (15.24% difference). By contrast, all the groups tended to NOT attribute these sentences to speakers living in Mexico (as indicated by the negative values in Figure 1), and again, this was related to group, with simultaneous bilinguals (group 1) showing the largest difference between the experimental and control stimuli for this group (−46.96% difference) and late bilinguals (group 4) showing the smallest difference (−11.43%). The participants in groups 3 and 4 also attributed these statements to people living in Spain, almost equally as to those who grew up in Southern California, but the participants in groups 1 and 2 did not. In summary, the simultaneous and early bilinguals strongly identified adapted borrowings as part of their own speech community and did not associate them with speakers living in Mexico, Spain, or another Spanish-speaking country. By contrast, the 1.5-generation and late bilinguals (living in Southern California) associated these phrases equally with bilinguals who grew up in Southern California or speakers in Spain and less so with people living in Mexico, but not as strongly as early bilinguals did.

Figure 1. Difference in response rates between experimental and control sentences for adapted borrowings separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

Figure 1. Difference in response rates between experimental and control sentences for adapted borrowings separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

8.2. Non-Adapted Borrowings

The pattern of responses was similar for non-adapted borrowings as that seen for adapted borrowings with a general preference across the groups for associating these forms with bilinguals who grew up in Southern California and not with people living in Mexico (see Figure 2). Non-adapted borrowings were mostly assigned to bilinguals living in Southern California (44.42% experimental vs. 18.60% control), bilinguals living in another part of the U.S. (18.37% experimental vs. 10.93% control), and people living in Spain (45.35% experimental vs. 32.09% control), but not to people living in Mexico (23.95% experimental vs. 57.67% control) or bilinguals who grew up in Mexico but live in Southern California (11.16% experimental vs. 24.19% control). The pattern of responses was similar for non-adapted borrowings as that seen for adapted borrowings (see Figure 2) with a general preference across the groups for associating these forms with bilinguals who grew up in Southern California and not with people living in Mexico. The preference for associating these forms with bilinguals in Southern California was more similar across the groups (ranging from 21.90% to 29.17%) than it was for adapted borrowings. The dissociation of these forms with people living in Mexico still showed a graded relationship across the groups with the simultaneous bilinguals showing the largest effect (−49.57%), but the 1.5-generation and late bilinguals showing smaller but more similar effects (−24.44% for group 3 and −25.71% for group 4). Another similarity between adapted and non-adapted borrowings was that only groups 3 and 4 associated these forms with speakers living in Spain. However, the late bilinguals showed a much stronger preference for identifying these forms with speakers in Spain (41.90%) compared to bilinguals who grew up in Southern California (21.90%), whereas the 1.5-generation bilinguals showed a more similar association with both of these groups (20.00% for Spain and 24.44% for bilinguals in SoCal). Another pattern that differed somewhat from that seen for adapted borrowings is that the participants in groups 1 and 2 showed a dispreference for associating these forms with bilinguals living in Southern California who grew up in Mexico, with the size of the effect more pronounced for the simultaneous bilinguals (−29.96%) than for the early sequential bilinguals (−16.67%). In summary, like adapted borrowings, non-adapted borrowings were associated with bilinguals who grew up in Southern California and not with speakers living in Mexico. However, the 1.5-generation and especially the late bilinguals also associated these forms with people living in Spain, and the simultaneous and early sequential bilinguals showed a negative association with bilinguals who grew up in Mexico but live in Southern California.

Figure 2. Difference in response rates between experimental and control sentences for non-adapted borrowings separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

Figure 2. Difference in response rates between experimental and control sentences for non-adapted borrowings separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

8.3. Caló Forms

Interestingly, Caló and patrimonial forms were not attributed to any of the groups more than the control sentences. However, Caló forms were assigned less often to people living in Spain (17.21% experimental vs. 34.42% control), people living in other Spanish-speaking countries (15.12% experimental vs. 32.33% control), and bilinguals living in Southern California (14.19% experimental vs. 20.93% control) or another part of the U.S. (6.05% experimental vs. 13.26% control). These forms elicited no preference by the participants in groups 1 and 2, but they did elicit negative associations with all three bilingual groups as well as speakers in Spain and in other Spanish-speaking countries (see Figure 3). By contrast, the 1.5-generation and late bilinguals showed a small association of these forms with speakers living in Mexico, and the late bilinguals showed an association of these forms with bilinguals in Southern California who grew up in Mexico. The late bilinguals also showed a negative association with speakers in Spain and other Spanish-speaking countries but not with the other bilingual groups. In sum, the late bilinguals associated Caló forms with their own speech community—bilinguals living in Southern California who grew up in Mexico—and to a lesser extent with speakers living in Mexico. The early bilinguals, however, did not associate Caló forms with this group, and in fact, showed no clear association of this form with any of the six groups that they were asked to choose from.

Figure 3. Difference in response rates between experimental and control sentences for Mexican Caló forms separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

Figure 3. Difference in response rates between experimental and control sentences for Mexican Caló forms separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

8.4. Patrimonial Forms

A similar pattern was observed for patrimonial forms as was seen for Caló with some exceptions (see Figure 4). Patrimonial forms were assigned less often to people living in Mexico (48.14% experimental vs. 59.07% control) and to all three bilingual groups (bilinguals who grew up in Mexico: 18.37% experimental vs. 23.95% control; bilinguals who grew up in Southern California: 17.67% experimental vs. 22.56% control; bilinguals who grew up in another part of the U.S.: 10.93% experimental vs. 13.49% control). The late bilinguals showed a small preference for associating these forms with their own speech community (bilinguals living in Southern California who grew up in Mexico) while none of the other groups did, and groups 1 and 2, in fact, showed a disassociation of patrimonial forms with this group. Another similarity across Caló and patrimonial forms is that the simultaneous bilinguals showed no association of these forms with any of the six options. However, unlike for the Caló forms, all the groups showed a negative association with speakers living in Mexico. Instead, the 1.5-generation and late bilinguals associated these forms with speakers living in Spain, and the early sequential and late bilinguals associated these forms with speakers living in other Spanish-speaking countries.

Figure 4. Difference in response rates between experimental and control sentences for patrimonial/rural forms separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

Figure 4. Difference in response rates between experimental and control sentences for patrimonial/rural forms separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

8.5. Central Mexican Colloquial Forms

A similar pattern was found for central Mexican colloquial forms as that seen for Caló (see Figure 5); colloquial forms were assigned less often to people living in Spain (17.67% experimental vs. 30.00% control), people living in other Spanish-speaking countries (16.98% experimental vs. 29.07% control), and bilinguals living in Southern California (13.95% experimental vs. 24.19% control) or another part of the U.S. (6.98% experimental vs. 13.72% control). However, unlike Caló forms, colloquial forms were assigned more often to people living in Mexico (64.65% experimental vs. 52.09% control) for all four groups (range of 5–20% difference) and were negatively associated with people living in Spain (range of −2 to −21% difference) or other Spanish-speaking countries (range of −5 to −19% difference). All four groups negatively associated these forms with bilinguals who grew up in Southern California, but this association was strongest for the simultaneous bilinguals (−17.39%) and grew smaller across the four groups (Group 2: −12.50%; Group 3: −6.67%; Group 4: −2.86%). Moreover, Groups 1–3 showed negative associations with bilinguals who grew up in a different part of the U.S. (range of −8% to −10% difference), whereas Group 4 did not. In sum, Central Mexican colloquial forms were associated with speakers living in Mexico but not with those who had moved to Southern California and were negatively associated with Spanish speakers in other Spanish-speaking countries, including Spain.

Figure 5. Difference in response rates between experimental and control sentences for Central Mexican colloquial forms separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

Figure 5. Difference in response rates between experimental and control sentences for Central Mexican colloquial forms separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

8.6. Redundancies

Redundancies were most often assigned to bilinguals from Southern California (32.33% experimental vs. 21.40% control) and less often to people living in Mexico (46.74% experimental vs. 55.58% control), Spain (26.98% experimental vs. 32.33% control), or another Spanish-speaking country (23.49% experimental vs. 33.26% control). All the groups associated redundancies with bilinguals who grew up in Southern California, but this association was strongest for the early sequential bilinguals (15.83% difference). All the groups showed a negative association with speakers living in Mexico (range of −4% to −16%), particularly for groups 3 and 4, and speakers living in other Spanish-speaking countries (range of −2 to −17% difference), particularly group 1 (see Figure 6). Interestingly, this is the only stimulus category that was associated most strongly with bilinguals who grew up in another part of the U.S., though only by groups 2–4 (most strongly by group 3; 13.33% difference). In addition, the late bilinguals associated these forms with speakers living in Spain (13.33% difference), whereas the other three groups showed a negative association of these forms with this group (range of −8 to −16% difference). In sum, redundancies tend to be associated most often with bilinguals who grew up in the U.S., either in Southern California or outside of that region, and not with speakers living in Mexico or other Spanish-speaking countries, with one exception (the late bilinguals associated this form with speakers in Spain).

Figure 6. Difference in response rates between experimental and control sentences for redundancies separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

Figure 6. Difference in response rates between experimental and control sentences for redundancies separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

8.7. Taboo Forms

Lastly, taboo forms were more often assigned to people living in Mexico (65.68% experimental vs. 60.23% control), bilinguals in Southern California who grew up in Mexico (28.84% experimental vs. 20.70% control), and bilinguals who grew up in Southern California (22.56% experimental vs. 13.95% control), and less often to people living in Spain (21.40% experimental vs. 33.02% control) or another Spanish-speaking country (16.51% experimental vs. 30.23% control). These taboo forms showed quite stark group differences (see Figure 7). All the groups positively associated these forms with speakers from Mexico who live in Southern California (range of 4–11% difference). However, only the simultaneous and early sequential bilinguals showed a clear association of this form with bilinguals who grew up in Southern California (Group 1: 13.91%; Group 2: 18.33%), while the 1.5-generation and late bilinguals did not. The early sequential bilinguals also associated these forms with bilinguals who grew up outside of Southern California (6.67% difference), but no other groups did so. Additionally, the early sequential and 1.5-generation bilinguals associated these forms with speakers living in Mexico (Group 2: 7.50%; Group 3: 11.11%). All four groups showed a negative association of these forms with speakers from other Spanish-speaking countries, and groups 1–3 also showed a negative association with speakers from Spain.

Figure 7. Difference in response rates between experimental and control sentences for tabú forms separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

Figure 7. Difference in response rates between experimental and control sentences for tabú forms separated by group and response. Positive values indicate higher response rates for experimental sentences than matched control sentences, while negative values indicate lower response rates for experimental sentences than matched control sentences.

9. Discussion

The results provide a preliminary window into the construction of the sociolinguistic realm for Spanish–English bilingual speakers in SoCal. While the results are based on the comparison of raw percentages and not drawn from statistical inferences, several clear and informative patterns emerged. A novel finding from this study is that not all Spanish–English bilingual speakers recognize these categories as characteristic of the language practices of SoCal bilinguals to the same extent. “Popular” Mexican Spanish is often categorized in the literature as stigmatized and described as “colloquial/informal”, “rural”, or “archaic”. However, employing a usage-based approach, we view these conventionalized structures as symbolic elements in SoCal bilinguals’ speech. The speakers in our study differentiated these categories of stigmatization, identifying some of them as more characteristic of SoCal Spanish compared to others. For example, the participants demonstrated an awareness of English borrowings, as they noticed a difference between the adapted and non-adapted contact forms. In this way, non-adapted borrowings are salient for speakers in SoCal as they stand out from more adapted ones, historically established in the Southwest region, and this is particularly evident for simultaneous bilinguals, who grew up in SoCal exposed to both languages from birth. Thus, by being salient, non-adapted borrowings, and to a lesser degree adapted ones, constitute exemplar tokens evoking emblematic performances from the speakers in this region (see Drager and Kirtley 2016). Similar to borrowings, redundancies seemed to be recognized by the respondents, particularly by the simultaneous bilinguals, as iconic salient elements, that is, as tokens of an identified SoCal variety. As a reminder, awareness was measured in our study as the difference in response rates for the experimental sentences compared to the control sentences in a given stimulus category with greater awareness reflected in larger difference scores (whether positive or negative). Salience was measured across the groups and response options so that the most salient forms were those associated with the participant groups’ own linguistic community. We found that the awareness regarding border Caló and patrimonial forms were similar to each other but different from that seen for borrowings. Both the Caló and patrimonial categories have been grouped as elements of a border Mexican and U.S. Southwest variety spoken since colonial times. Patrimonial categories have commonly been labeled as “rural” or “archaic” forms that are still used in the Americas, including in urban areas, and prevail in SoCal. Thus, the lack of differentiation between these forms and the more standardized (and thus, generalized) forms presented in the control phrases may reflect that they are perceived as part of speakers’ repertoire and that there is a lack of salience (noticing) in the way in which they are perceived and therefore, a lack of awareness of them. Thus, the participants may perceive these non-standardized, stigmatized forms as variants and not as salient “errors” (see Squires 2016). On the other hand, the forms from the Central Mexican Colloquial and taboo categories were more salient and perceived within the awareness realm of SoCal bilinguals as used by people in Mexico. In this manner, these forms would represent exemplars of “Mexican Spanish” for SoCal bilinguals.

Bybee (2013) observes that “it is repetition that leads to conventionalization of categories and associations, as well as to the automation of sequences” (p. 50). We chose highly conventionalized lexical items within the categories of stigmatization that have been largely documented and marginalized in Hispanic linguistics and as being part of the so-called Mexican-based vernacular “Spanish from Southwest region in the U.S., rural or popular varieties, border Caló, among other labels” (e.g., Aguilar Melantzón et al. 1985; Alarcón 1978; Espinosa 1915; Lamar Prieto 2014; Lipski 2011; Moreno de Alba 2003; see also “El Español de los Estados Unidos” by Escobar and Potowski 2015). Thus, our study centers not on demonstrating the frequency of usage of these stigmatized forms from Mexican varieties, but on the incorporation of them as exemplars of this imaginary of Mexican language practices and realm, as exemplars that are already part of speakers’ sociolinguistic practices and habitus. While most of the existing studies in Hispanic linguistics have distinguished some of these categories of stigmatization (e.g., English borrowings vs. patrimonial or archaic forms), our study is innovative in the sense that it also demonstrates the differences in perceptions and attitudinal stances, and the saliency (or not) of different categories of these frequent forms through different generations of speakers in the U.S.

These results, together with those from the same sample reported elsewhere (Holguín Mendoza et al. in press), expose the respondents’ different levels of awareness of the stylistic variants traditionally grouped as “U.S. Spanish”. Our results also reveal that there are differences in awareness depending on the age at which bilinguals came to inhabit SoCal’s linguistic milieu. In this manner, the mental spaces, that is, cultural chronotopes, comprising SoCal speakers’ sociolinguistic awareness, are continuously being reimagined as they become nourished by racist, classist, and other sociolinguistic ideologies that are internalized within a complexity of differentiation processes. As seen with these results, elements from Mexican Spanish that either have been brought over or innovated since colonial times, interact in SoCal under new conditions in the U.S. More importantly, we have observed how all these elements of SoCal Spanish, even though they can be historically traced to Mexican Spanish and language in contact on the borderlands, have been semiotically and ideologically reformulated by their users. Some of these forms seem to have been integrated into their repertoires and are not as salient; others have become tokens. These semiotic reformulations greatly differ from sociolinguistic ideological configurations in Mexico and other Spanish-speaking centers, including the very same communities in the U.S. (Holguín Mendoza et al. in press), but not less importantly by Spanish researchers and educators who often consider these forms as “incorrect” or “informal”. Given that this is one of the first studies, to our knowledge, to explore the perceptions of informal/stigmatized language forms among different groups of Spanish–English bilingual residents, particularly as they relate to the respondents’ own versus others’ language practices, and given that the results were drawn from descriptive rather than inferential statistical analysis, we consider the findings presented here to be preliminary. Future research could explore these questions further by including survey questions that involve single-response questions as opposed to multiple-response questions so that inferential statistics can be applied to the data analysis. Moreover, follow-up qualitative studies could explore the language attitudes of various populations of Spanish–English bilingual residents of Southern California, such as those surveyed here, including how they characterize and use various forms of language and how those practices reflect a unique sociolinguistic community.

10. Conclusions

These results highlight the importance of revising how stigmatized categories are incorporated into research and taught in language classrooms as a normalized myth of what U.S. Spanish is, considering that the SoCal community in the United States employs sociopragmatic functions and values that differ from other Spanish-speaking groups. Language practitioners must recognize these complex language practices from the communities and incorporate them into their research and pedagogical approaches, particularly in teaching Spanish to Latinx bilingual speakers. Our results also underscore the need to incorporate into our language curriculum development discussions of how communities and their sociolinguistic practices become iconized and invisibilized and how speakers across social groups continue to reproduce recursively ideologies of sociolinguistic marginalization (see Valdés and Parra 2018; see also Martínez 2006). Thus, there is an increasing need to incorporate critical pedagogical approaches that address differences and power toward and between speakers of Spanish, such as those speaking varieties of Greater Mexico and the borderlands. Many Spanish-language textbooks in the U.S. still omit or disregard these various linguistic categories as “non-standard”, completely ignoring and marginalizing local students’ identity formations, perpetuating marginalization in the classroom and society.

Finally, we uphold the analysis of critical complexity in language research and applied linguistics, such as usage-based linguistics that includes self-reflexivity for both our research designs and theories to interrogate academic discourses that persist in maintaining bilingual speakers on the margins of research and pedagogy (Higby et al. 2023; Holguín Mendoza et al. in press). We advocate following recent work in linguistics and sociopragmatics that increasingly focuses on the complexities of social interaction (see Amorós-Negre et al. 2021; Haugh et al. 2021) to develop critical awareness and the literacy of discriminatory and oppressive social hierarchies embedded in language and discourse. As researchers and educators dedicated to sociolinguistic justice, the first steps in accomplishing these objectives are to thoroughly investigate these prejudices and language ideologies related to race and class (among other aspects) and reassess how these beliefs influence our attachment to social hierarchies; this paradigm shift must continue in linguistics since the costs are too great to ignore.