Latino Studies

, Volume 16, Issue 1, pp 65–90 | Cite as

The social meaning of Spanish in Miami: Dialect perceptions and implications for socioeconomic class, income, and employment

  • Phillip M. Carter
  • Salvatore Callesano
Original Article


In this study, we investigate the ways in which three national-origin varieties of Spanish commonly heard in metropolitan Miami—Cuban, Colombian, and Peninsular—are conceptualized by young adult residents of Miami-Dade County in terms of implicit perception. Further, we test whether or not perceptions about Spanish can predict social outcomes in the domains of labor, employment, and income. Three male residents of Miami were asked to read a text in their home variety of Spanish. All men were college educated in their respective countries of origin (Cuba, Colombia, and Spain) and are professionally employed in Miami. For each voice heard, participants were given background information about the speaker, including the parents’ country of origin. In some cases, the parents’ national-origin label matched the country of origin of the speaker (Speaker: Cuba, Origin-label: Cuba), but in other cases, the background information and voices were mismatched (Speaker: Cuba, Origin-label: Spain). This manipulation allows us to separate the perceptions based on the elements of the speech signal from the provided social information. Participants were asked to rate the voice/background permutations on five-point Likert-scales for a range of personal characteristics. Data were analyzed for significance using a three (dialect) X four (label) within-subjects ANOVA with a series of specific statistical contrasts. Our analyses of these judgments showed three kinds of significant effects: (1) main effects of language variety, (2) main effects of the background label, and (3) interaction effects. Overall, we find that adolescent Latin@s in Miami-Dade exhibit divergent perceptions of national-origin varieties of Spanish and that they use sociolinguistic differences to make predictive judgments about nonlinguistic, social attributes related to socioeconomic class, including family wealth, personal income, and profession.


Miami Spanish Dialect perception Sociolinguistics Cuban Spanish Latinidad 

The social meaning of language diversity in US Latin@ communities

Popular figurations of the social category “Hispanic/Latin@” take US Latin@s to be “Spanish-speaking” or “bilingual.” These characterizations are, of course, both limited and limiting and ignore the immense linguistic diversity of language use in US Latin@ communities. This diversity is characterized by five sociolinguistic phenomena. First, US Latin@s speak a diverse range of national-origin, regional, ethnic, and class-based varieties of Spanish (Escobar and Potowski 2015). Some of these varieties have a long-standing historical presence in the United States, while others reflect more recent patterns of immigration. Second, some of these varieties of Spanish may, over time, be influenced by sustained contact with English or, in the case of urban settings such as New York City, sustained contact with other varieties of Spanish (Otheguy and Zentella 2012). These situations of language contact may result in unique, localized varieties of “US Spanish.” Third, the interdisciplinary literature documenting the use of Spanish in the United States shows a clear and consistent pattern of language shift from Spanish to English across generations (e.g., Otheguy et al. 2000; Veltman 1988; Zentella 1997). The phenomenon of language shift in US Latin@ communities results in wide-ranging patterns of bilingualism and complex sociolinguistic patterns of language use. Fourth, long periods of sustained contact between Spanish and English have given rise to the bilingual mixed variety known as Spanglish (Carter and Lynch 2015), which some US Latin@s claim instead of Spanish (Fuller 2012; Zentella 1997). Finally, US Latin@s speak a wide range of English-language varieties, which differ by region, national-origin group, socioeconomic class, and racio-ethnic identity and other forms of social affiliation (Fought 2002; Wolfram et al. 2004; Mendoza-Denton 2008; Carter 2013). In short, the languages and language varieties of US Latin@s are as diverse as the bodies, national-origin groups, and nonlinguistic sociocultural practices that characterize Latin@ experiences in the United States. Despite the immense diversity that characterizes language in US Latin@ communities, however, interdisciplinary scholarship seeking to understand and unpack the differences consolidated by the term “Latin@” rarely pauses to consider the role that linguistic difference plays in the lives of US Latin@s, either in terms of individual and group identity or in terms of social inequality. This is surprising in light of the long history of scholarship in linguistic anthropology, sociolinguistics, discourse analysis, and allied disciplines that demonstrate the fundamental role that language plays in constructing social difference (Bourdieu 1991; Blom and Gumperz 1972; Eckert 2000; Mendoza-Denton 2008) and promoting forms of educational, socioeconomic, sociocultural, and legal inequality (Lippi-Green 1997; Zentella 1997, 2014). In other words, linguistic diversity in US Latin@ communities is not merely an empirical reality to be accounted for by linguists and demographers, but rather a vital symbolic resource used in the making of diverse Latin@ sociocultural identities with attested links to material, sociological outcomes.

In Miami—the United States’ most Latin@ and most Spanish-speaking metropolitan region—Spanish is a potentially important unifying force, as set forth in the introduction to this special edition. It unites individuals from Miami-Dade’s diverse national-origin groups, Cubans with Colombians, Venezuelans with Nicaraguans, Peruvians with Spaniards, and so forth. Given the majoritarian status of Miami-Dade’s Latin@ population in numerical terms, Spanish is fairly ubiquitous. Spanish in Miami City, with a population that is 57.3% foreign-born (US Census Bureau 2010), also connects recent immigrants with the more locally established. It unites Miamians with their families’ countries of origin and figures as an important unifying force in local politics. But just as much as Spanish unites, it also differentiates. As Carter and Lynch (2015) argue, Miami-Dade is now the most dialectally diverse Spanish-speaking metropolitan area in the world. The national-origin varieties of Spanish, and the racio-ethnic, class, and regional differences within them, are a resource for constructing and reinforcing difference in national origin, socioeconomic class, and racio-ethnic identity, as well as in constructing differences in immigration status, such as generation of arrival, means of arrival (e.g., airplane, boat, raft), and “new” versus “old” arrival. For example, an important part of “being” Cuban in Miami-Dade is speaking a variety of Spanish that indexes that national-origin formation, which is different from the national-origin varieties that index Colombian, Nicaraguan, and Venezuelan identities, and so on. At the same time, Cuban Spanish in Miami-Dade is also stratified by nonlinguistic social factors, with certain linguistic structures—such as the syllable-final aspiration of s (loh carroh for los carros)—serving as semiotic links to classed and racialized identity formations (Lynch 2009). To have arrived in Miami-Dade in the 1990s or 2000s via raft (balseros) does not convey the same status as having arrived via boatlift in the 1980s (Marielitos), which in turn does not convey the same status as having arrived in the 1960s or 1970s via airplane. Though all of the people inhabiting these groups speak Spanish, and although they all speak a variety that can be identified as “Cuban,” differences in speech nevertheless exist among these groups, and these differences can become symbolic or indexical markers (Silverstein 2003) of social status. Although work by linguistic anthropologists and sociolinguists show that US Latin@s actively construct “Latino” as a discrete racial category that contrasts with “Black,” “White,” and “Asian” (e.g., Slomanson and Newman 2004) we must also acknowledge the ways in which inequality—including racial inequality—lives within the ethno-racial category “Latino.” In this article, we argue that differences in language play a fundamental, not ancillary, role in constructing the kinds of difference within the category “Latino” explored throughout this volume, and by Latino studies scholars working in a variety of disciplinary formations.

In the study we present here, we are concerned with the ways in which three national-origin varieties of Spanish commonly heard in metropolitan Miami—Cuban, Colombian, and Peninsular—are perceived by residents. Further, we are interested in understanding the extent to which perceptions about Spanish can predict social outcomes in the domains of labor, employment, and income. In this regard we are primarily interested in perception of Spanish-language variation as it pertains to issues of socioeconomic class, though we appreciate that in this context class is imbricated with racio-ethnic identity, national origin, and other vectors of identification. In order to test Miamians’ perceptions of Spanish-language difference, we have turned to the methods in the field of perceptual dialectology (Bucholtz et al. 2007; Fridland et al. 2004; Preston 1989; Niedzielski and Preston 2000, inter alia) and social psychology (Lambert et al. 1960), which allow us to explore perceptions of language and their social correlates at the implicit level. We describe these methods in detail later in the article. In the following section, we lay out the relationship between language variation, social stratification, and ideology, before turning to an overview of the sociolinguistic situation in Miami.

Linguistic variation, ideology, and social stratification

In 1966, sociolinguist William Labov conducted a large-scale linguistic analysis of the speech of New Yorkers from various socioeconomic class groups. In contradistinction to the linguistic theory that preceded his work, Labov found that the variable use of linguistic structures was not the result of “free variation,” in which speakers randomly or carelessly shuttled between linguistic forms, but rather the result of naturally occurring patterns of language variation that were stratified by social class divisions. He showed that the speech of New Yorkers was systematic, patterned, and rule-governed, not only for speakers at the top and in the middle of the socioeconomic class hierarchy, but also for lower- and working-class speakers at the bottom. Nonstandard or “vernacular” speech, it turned out, was just as systematically structured as speech considered “standard.” For all speakers, variation in the linguistic system was governed by both linguistic and social constraints, corresponding to meaningful divisions in social structure. For example, while New Yorkers sometimes produced the nonstandard form [n] for [ŋ] (i.e., swimmin’ for swimming), this usage was also stratified by socioeconomic class, with upper-middle-class speakers using the nonstandard form the least, and low-income speakers using it the most. Labov concluded that dialect differences were “an element in a highly systematic structure of social and stylistic stratification” (1966, p. vii). In the five decades that have passed since the publication of Labov’s study in New York City, similar patterns of sociolinguistic stratification have been attested time and again in speech communities across the United States and around the world, not only in English but in a range of typologically related and unrelated languages, including Spanish (e.g., Shuy et al. 1968; Cedergren 1973; Trudgill 1974; Lavandera 1978; Modaressi 1978; Feagin 1979; Naro 1981; Poplack 1980; Horvath 1985). As a result of this work, the correlational relationships among identity categories and social positions (e.g., socioeconomic class) and linguistic variation is very clear. Eckert (2005, p. 2) notes that these kinds of sociolinguistic studies “established a regular and replicable pattern of socioeconomic stratification of variables, in which the use of non-standard, and geographically and ethnically distinctive variants, correlate inversely with socioeconomic status.” Linguistic variation—including the sort that produces perceptual differences in the “dialects” of a particular language—is itself a fundamental part of the linguistic system (Guy 2005).

We begin this section with a summary of the work of quantitative sociolinguistics in order to set forth a basic yet widely misunderstood point about language: all human language systems are systematic, patterned, and rule-governed in their own right, irrespective of how they are evaluated socially. However, because some patterns of sociolinguistic variation correspond to social differences in racio-ethnic identity, socioeconomic class, region, and other vectors of social difference, it is easy to fold differences in language—or perceived differences in language—into ideological formations about the social groups from which the linguistic difference emerges or is thought to emerge. In other words, stereotypes about a social group or speech community become the stereotypes associated with their language or language variety (Wolfram and Schilling 2015). Bourdieu (1984, 1991) theorized that standard language varieties—most often based on the dialect forms associated with social elites—get constructed in opposition to nonstandard varieties, and over time the linguistic preferences and patterns of usage of the elites are normalized and constructed as “natural” or “legitimate,” while the forms of speech associated with non-elites are constructed as naturally and obviously inferior. In Bourdieu’s view, language in late modernity is thus a form of symbolic capital that reinforces an ostensibly “natural” or “underlying” social structure in which dominant groups can effortlessly reassert their dominance, while marginal groups are systematically re-marginalized. Sociolinguist and social theorist Rosina Lippi-Green notes that widespread beliefs about the putative superiority of certain language varieties and the corresponding inferiority of others are ideological in nature. Standard language ideology, as she terms it, is “bias toward an abstracted, idealized, homogenous spoken language, which is imposed and maintained by dominant bloc institutions and which names as its model written language, but which is drawn primarily from the spoken language of the upper middle class” (1997, p. 64). Standard language ideology plays out with systematic material consequences in the educational system, such that children whose home dialect corresponds to the school standard are automatically rendered exceptional, gifted, and prepared for school, while children who speak nonstandard dialects in the home are automatically rendered remedial, deficient, or unprepared for school. In legal and judicial contexts (e.g., Rickford and King 2016), speakers of standard varieties are considered more reliable and trustworthy, while speakers of nonstandard varieties are considered less reliable and less trustworthy. In a number of experiments focused on linguistic profiling (Baugh 2003) in residential and housing contexts, callers inquiring about the availability of apartments have been shown to be denied housing or asked to pay larger deposits or higher rents when using nonstandard varieties than when they made the same inquiry using standardized English (Purnell et al. 1999; Massey and Lundy 2001). Thus, the material consequences of ideological beliefs about language in these contexts—from lowered grades and test scores to lost court cases—are not inconsequential. Though the specific linguistic features vary from place to place and from language to language, sociolinguists studying the indexical link between linguistic feature and social status (Babel 2014; Bradac and Wisegarver 1984; Docherty et al. 2013; Fridland and Bartlett 2006) have shown that the evaluative process is mostly the same: the features associated with prestige groups, including the wealthy and the highly educated, are perceived positively, and the use of those features by those groups thus reinforces their prestige status.

The epistemology of language in which spoken forms can be described in terms of binaries such as “good/bad” or “legitimate/illegitimate” is tightly bound to questions of nation and nationalism because, as Benedict Anderson (1983) shows, the very idea of standardization arose alongside the development of the nation-state. Within the European model of nation-state development, states are constructed as inherently or naturally monolingual and monoethnic. Language remains a fundamental symbolic resource in nation-building efforts, and national languages, particularly the constructed “standard” varieties, are therefore tightly bound up with national identity the world over (Tetel Andresen and Carter 2016). Monolingual dictionaries, national language academies, and systems of education reinforce the ostensibly natural link between geographic location, language, and ethnoracial groups.

In the case of Spanish, which is spoken in over twenty countries in Europe, North America, and South America, national varieties of the language are considered distinctive and representative of national identity. These language varieties—and the language ideologies that attend them—are closely tended to by the Real Academia de la Lengua Española (Royal Academy of the Spanish Language) and by the twenty-one other national academies that make up the Association of Spanish Language Academies. A crucial point is that the national varieties of Spanish are themselves not constructed as socially neutral. For example, on 16 February 2015, the director of La Real Academia Española, Dario Villanueva, gave a radio interview to the hosts of Uniminuto Radio, a station based in Bogota, Colombia. One of the hosts asked for Villanueva’s assessment of Spanish in Colombia. He replied by saying, “Los colombianos hablan tan bien el español que sorprende positivamente la habilidad con que se expresan y eso nos estimula a mejorar y a darnos cuenta de los defectos y errores que nosotros mismos podamos cometer.” [Colombians speak Spanish so well that the ability with which they express themselves pleasantly surprises us, and that inspires us to get better and recognize our own defects and errors that we ourselves may make.] A widespread belief in the Spanish-speaking world is that Colombian Spanish is among the best, most pure, or most refined varieties of Spanish, at least as it is spoken in the highland dialect region in cities such as Bogota. Viallanueva’s comments thus reified “Colombian Spanish” as a linguistic object that can be named as such, while reinforcing and propagating a social-ideological belief about Spanish already in circulation.

It is no coincidence that Colombia, like Argentina, Chile, and Uruguay, is considered to be more middle class and more ethnically European, compared with other Spanish-speaking countries in Latin America. In contrast, it is often remarked that speakers of varieties of Spanish from the Caribbean and coastal areas of Central and South America “se tragan la ese [they swallow the s]” and “no pronuncian bien las consonantes [they don’t pronounce the consonants well].”1 These turns of phrase are not merely folk descriptions of the phonological processes that characterize socially neutral varieties of Spanish, but rather serve as oblique commentaries about the quality of these varieties and, by extension, the people who speak them. It is thus easy to formulate the erroneous idea that intelligent people pronounce—s, unintelligent ones do not. The semiotic link between phonological feature (e.g., /s/) and social trait (e.g., perceived intelligence) is thus reinforced as “real,” and the broader epistemological framework in which language serves as a proxy for ideologies about nation/race/class is perpetuated. As a result, regular patterns of language variation that produce as their effects the recognizable social and regional varieties of Spanish become part of the widespread ideological narrative about Spanish, its dialects, its speakers, and their relative value in terms of goodness, correctness, and other forms of social valuation. Here again, it is no wonder that the most stigmatized varieties of Spanish are those associated with Latin American countries with large indigenous, Black, and poor populations. It is likewise no wonder that the moneyed elites in those countries tend to avoid using the speech forms that link semiotically to those groups.

Linguists studying attitudes and perceptions of Spanish dialects have repeatedly found that Caribbean varieties are perceived negatively compared with non-Caribbean varieties (García et al. 1988; Alfaraz 2002; Otheguy et al. 2007). A crucial point is that speakers of Caribbean varieties often share these perceptions and reproduce stigmatizing attitudes about their home dialects after immigrating to the United States (Duany 1998). Suárez Büdenbender (2013) has found that Puerto Ricans were able to disambiguate Puerto Rican Spanish from Dominican Spanish in a controlled experiment and that they were able to use dialect recognition to make assessments about a speaker’s socioeconomic and educational background. In other words, stereotypes and attitudes about Spanish are not “merely” discursive, but are also woven into the complex mental representations speakers have about themselves and each other. Recent studies in perceptual dialectology (Mojica de León 2014; Quesada Pacheco 2015; Sobrino Triana et al. 2014) demonstrate that negative perceptions of varieties of Caribbean Spanish are robust in Latin America, including within the Caribbean.

We end our conversation about sociolinguistic variation, language ideology, and socioeconomic class in the context of Latin America, not only because our study focuses on Spanish, but also as an acknowledgement that many of the ideological beliefs about Spanish found in Miami-Dade are in a dialectical relationship with the broader ideological apparatus about the Spanish language, linked as it is to colonialism, European exceptionalism, and the kinds of racio-ethnic, classed, and place-based identities discussed in the introduction and throughout this volume. In the following section, we offer a short overview of the sociolinguistic landscape in Miami as it pertains to Spanish.

Spanish in Miami

As set forth in the introduction to this special edition, Miami-Dade is now a majority Latin@ metro region. As of the 2010 US Census, Miami-Dade was 65% Hispanic/Latin@, while Miami City was 70% Hispanic/Latin@. Table 1 provides data from the 2010 US Census on ethnicity, language use, and median household income for several municipalities in Miami-Dade County. Carter and Lynch (2015) make three points about these data that are important for the discussion about Spanish in Miami-Dade we are staging in this article. First, all of the major municipalities within Miami-Dade have populations that are at least 50% Latin@. Second, the use of LOTES (“languages other than English”) is high across all Miami-Dade municipalities. Although Haitian Creole, French, Portuguese, and Russian are important languages of the region, with sizeable numbers of speakers, we can safely assume that the most widely spoken language other than English spoken in the home is Spanish. Finally, when we compare the “percent non-English at home” with median household income, we see high percentages of Spanish not only in working-class areas such as Hialeah (median income: $31,648; % non-English: 94.2%), but also in middle-class areas such as Doral (median income: $69,300; % non-English: 88.8%) and in affluent areas such as Key Biscayne (median income: $104,554; % non-English: 79.9%). Thus, unlike other parts of the United States with large Latin@ populations, where Spanish is mostly a language of non-elites, Spanish in metropolitan Miami lives throughout the socioeconomic hierarchy. Given that Spanish is not only a language of the working class in Miami-Dade, we might predict that structural differences in the language indexing different national origin, racioethnic, and socioeconomic class differences will take on additional ideological weight when compared with situations where Spanish is already always understood to be a language of the poor and thus not relevant for making fine-grained socioeconomic class distinctions.
Table 1

Ethnic identity, language use, and income data for Miami-Dade County (US Census Bureau 2010)


% Hispanic/Latino

% White non-Hispanic

% Black

% Foreign-born

% Speak non-English language at home (age 5 +)

% High school degree (age 25 +)

% Bachelor’s degree (age 25 +)

Median household income

Miami-Dade Co.









Miami (city)









Coral Gables


















Miami Lakes


















Miami Beach


















Key Biscayne









In addition to its use across the socioeconomic hierarchy, Spanish in Miami-Dade differs from the national context in several other respects. At the national level, the largest national-origin group is Mexicans, who constitute 64.5% of the US Latin@ population (Lopez et al. 2013). This trend is reflected in the major cities of Texas, the Southwest, and the West. Mexican subpopulations make up the majority of the Latin@ population in eight of the ten most Latin@ cities in the United States. In contrast, the Mexican-origin population in Miami-Dade is relatively small (3%) compared with the Caribbean groups who make up Miami-Dade’s majority Latin@ population. Cubans (54%), Puerto Ricans (6%) and Dominicans (4%) make up about 65% of Miami Dade’s Latin@ population (Brown and Lopez 2013). Therefore, Spanish in the area is—numerically speaking—mostly Caribbean. Despite its strong Caribbean presence, Miami-Dade is now a hub for political exiles from across Latin America, and as such, the county’s Latin@ population is diversifying, as discussed in the introduction to this volume. Every large Spanish-speaking national-origin group is represented in Miami-Dade’s contemporary sociolinguistic landscape. In addition to the Caribbean groups already mentioned, Spaniards, Colombians, Venezuelans, Nicaraguans, Peruvians, and Argentines have all moved to the region is sizeable numbers and have brought with them diverse varieties of Spanish, making Miami-Dade one of the most dialect-rich Spanish-speaking places in the world.

Studying dialect perception

The study presented here is conducted in the tradition of perceptual dialectology developed in the United States by Preston (1993, 1996, 1999, 2002) and the social psychological study of language attitude research (Lambert et al. 1960; Niedzielski 1999). Unlike dialectology, the branch of linguistics concerned primarily with mapping regional language variation, perceptual dialectology and language attitude research is interested in non-experts’ beliefs and attitudes about language. Linguists and psychologists have devised a number of techniques for eliciting language attitudes, including asking participants to listen to and evaluate speech samples, draw labels on maps to identify “correct” or “pleasant” speech (Preston 1989), or to sort cards bearing the names of geographical regions into groups (Tamasi 2003). Alfaraz (2002) used an attitude questionnaire to study perceptions of Spanish dialects among Cuban Americans in Miami-Dade. In her study, participants rated the correctness and pleasantness of various Latin American varieties of Spanish, a variety referred to as “Peninsular Spanish” and two varieties of Cuban Spanish, one corresponding to Spanish on the island before the end of the Cuban Revolution in 1959, one corresponding to post-Revolution Cuban Spanish. Participants rated the Peninsular variety the highest on both the “correctness” and “pleasantness” scales, followed by pre-Revolutionary Cuban Spanish (the variety used by the original Cuban exiles to Miami). Post-Revolutionary Cuban Spanish (the variety used after the elites left) was ranked toward the bottom, followed only by that of Nicaragua, Mexico, Honduras, Puerto Rico, and Dominican Republic.

In our study of perceptions of Spanish in metropolitan Miami, we were more interested in capturing people’s implicit reactions to Spanish-language variation than with their explicit attitudes, which we have a good sense of from Alfaraz’s (2002, 2014) work. As such, we designed a perception experiment in which participants listened to three varieties of Spanish commonly heard in Miami-Dade, but were told they would be using their intuitions “to make guesses about strangers,” rather than make assessments about language varieties as such. Participants began the study by reading the following introductory prompt:

Recent scientific studies have shown that people can be amazingly good at guessing a stranger’s occupation, even by something as simple as seeing a photograph of the stranger’s bedroom, or seeing a sample of their handwriting. One study recently published in the journal Psychological Science found that people were about 65% accurate in judging a stranger’s occupation from a list of four options, just after hearing the person speak for 30 seconds.

The prompt was constructed in order to direct attention away from language while at the same time orienting participants to the idea that they could make accurate snap judgments based on short stimuli. Participants were then given four additional pieces of information:
  1. (1)

    all of the people you hear live in Miami;

  2. (2)

    all of the people you will hear will speak Spanish;

  3. (3)

    don’t worry if you do not speak Spanish yourself. Past scientific research shows you can make accurate intuitive judgments about people from hearing them speak even if you don’t know the language; and

  4. (4)

    each person you hear will be between the ages of thirty and thirty-two.

Miami was listed as the current place of residence for all speakers in order to ground the study and listener perceptions in the local sociolinguistic environment. We included the (false) information provided in point 3 in order to enfranchise non-Spanish-speakers in the study, since we were also interested in their judgments. Point 4 was included to control for perceptions of age.
Participants then encountered a brief profile for each voice they heard, which contained three pieces of information. The first piece of information was the same for all profiles: “Speaker lives in Miami.” We began each profile with this line in order to reorient participants to the local Miami-Dade context before listening. The second line of the profile contained an irrelevant piece of “filler” information, the speaker’s fictional birthday. The third line of the profile contained information about the speaker’s parents’ country of origin. This information was the experimental manipulation. We chose parents’ country of origin for two reasons. First, it recognizes the sociological reality in the county in which people are both mobile and of diverse national-origin heritage. Second, it allows us to explore how national-origin labels interact with speech to shape perceptions about Spanish. Four versions of the speaker profile were created, including three versions in which the speaker’s parents were said to have come from Spain, Colombia, or Cuba, plus one null-version for which family background information was not provided. These profiles were randomly assigned to the voices included in the study, yielding some “matches” (e.g., Cuban voice, parents from Cuba) and some “mismatches” (e.g., Cuban voice, parents from Spain). For example, participants might hear a Colombian voice, but believe the speaker’s parents were Cuban, a Cuban voice with Spanish parents, and so on. Each participant heard only three voices, with randomly assigned profiles, and no participant heard the same voice more than once. All voice-profile permutations were distributed evenly throughout the participant sample, yielding a robust number of responses per permutation. Those cases in which parental background information was omitted represented a pure perceptual dialectology condition, with no experimental manipulation. Table 2 shows the visual presentation of this information in the survey, which appeared on the screen as participants listened to each sample.
Table 2

Speaker profile

Speaker lives in Miami

Born on July 9

Parents are from Cuba

Testing implicit perceptions: Language varieties, instrument, and participants

The voices used as the instrument in the study come from recordings made with three male residents of Miami-Dade who are originally from Barcelona (Spain), Bogota (Colombia), and Havana (Cuba). All speakers were college educated in their country of origin, were between the ages of twenty-five and thirty-five, are currently and professionally employed in Miami-Dade, and have lived in the United States for at least one year. Each of the three speakers was asked to read the following brief text, borrowed from Carter and Lynch’s (2014) matched-guise experiment.

Es increíble como todavía las compañías de cigarrillos gastan billones de dólares cada año para promover el consumo de este producto. Es de conocimiento general que el fumar y usar tabaco causan cáncer y enfermedades del corazón, pero en el caso de los niños es más difícil que tomen conciencia acerca de este riesgo, ya que no entienden que hay enfermedades que pueden contraer al largo plazo.

Digital recordings of each speaker were made using a ZOOM H1 handheld audio recorder. Sound files were edited in the acoustic phonetics software program PRAAT (Boersma and Weenink 2013) in order to remove pauses and other disfluencies. Finally, each recording was cut down to a similar length (25 seconds).

A limitation of this type of research—and, indeed, of this study—is that although the different voices used to embody the national-origin varieties tested are reading the same passage, the voices themselves still differ from one another in terms of voice quality, pitch, intonation and various other aspects related to the performance of reading. One way of addressing this issue is for participants to hear several voices representing each dialect group being tested. In general, this method is untenable, in that it adds a prohibitive amount of time to the test situation, significantly increasing the rate of participant attrition. This problem would be especially pronounced in the current study, given the already high number of conditions produced by crossing each dialect group with several national-origin background labels.

We must also acknowledge that the dialect terms we use in this article—Cuban, Colombian, and Peninsular—are problematic, in that they imply that the voices participants heard are representative of dialect regions that correspond to national boundaries. We do not wish to reinforce that misunderstanding here. In addition to the socioeconomic-based differences described earlier, Spanish-language varieties also vary by geographic region, and it must be noted that dialect boundaries rarely correspond to national borders. Spanish in Cuba, Colombia, and Spain are of course geographically heterogeneous, and we use the terms “Cuban,” “Colombian,” and “Peninsular” only as shorthand for the varieties used in our experiment, and in recognition that these varieties are socially constructed as belonging to the nation-states in which they are spoken.

Participants, mostly undergraduate college students from Florida International University in Miami, were recruited during a two-week period during the fall 2014 semester to take part in the study, which was programmed and administered online using Qualtrics, an online survey platform used by social scientists that facilitates online data collection. In all, 292 participants completed the study. Of the participants, 67% (n = 195) identified as Hispanic/Latin@,2 and 33% (n = 96) identified as non-Latin@, a category that includes Anglo Whites,3 African Americans, and other ethnic groups. The participant profile reflects the demographics of Miami-Dade County and of the university where the sample was collected. In terms of language background, 78% of the participants (n = 228) said they were “native speakers” of English, and 52% (n = 152) said they were “native speakers” of Spanish. We included non-Spanish-speaking participants in the aggregate sample in order to study the extent to which they are able to discriminate among and make assessments about Spanish-language varieties commonly heard in Miami. Though this decision may be untenable in other regional contexts in the United States, we contend that including all ethnolinguistic groups in the study design makes sense in Miami, where everyone grows up hearing Spanish, even those who do not speak it.

Participants made twenty-six judgments for each voice/label combination they encountered in the survey. These judgments correspond to five general themes: personality characteristics, patterns of language use, employment, income, and family background. As our focus here is on socioeconomic class, this paper reports on the data for employment, income, and family background, yielding a total of twelve judgments. Judgments about employment and family background were made using five-point Likert scales, while assessments of income were made by selecting hypothesized annual salaries from a drop-down list. For the question about employment, participants were asked how likely it was that the speaker worked (a) as a salesperson at a cell phone store, (b) behind the counter at a local coffee shop, (c) as the office manager at a medical supplies company, (d) as an executive at a marketing firm, or (e) as an attorney. Participants were also asked how likely it was that the speaker comes from a family that (a) values hard work, (b) gives lots of opportunities to get ahead in life, (c) invests a lot in education, (d) is pretty poor, and (e) the previous generation did not have much choice about what they would do for a job. Demographic information on the participants was collected at the end of the experiment.

How Miamians perceive Spanish dialect variation

In this section we describe the results for the perception experiment, focusing first on general perceptual patterns in the aggregate, with Latin@ and non-Latin@ participants considered together, and then only on the perceptions of the Latin@ participants.

For our statistical analysis, we conducted a three (dialect) X four (label) within-subjects ANOVA with a series of specific statistical contrasts. Because of the planned missing data—not every speaker heard every three X four permutation4—we estimated the ANOVA and contrasts using a twelve-group model, freely estimating means and variances of the outcome, in the structural equation modeling software Mplus v7.4. These models were estimated separately for each of the seventeen outcomes. Mplus accommodated the missing data (which were Missing Completely at Random) using a direct maximum-likelihood estimation from the raw data. This estimator results in unbiased estimates and retains maximal statistical power. Standard errors were computed using a sandwich estimator (“MLR” in Mplus). Tests were computed using a series of model constraints (e.g., equality of marginal means for the main effects) and a scaled likelihood-ratio Chi squared statistic for the fit of the constrained model relative to the perfectly fitting unconstrained model. Where a post hoc correction was used, we applied the Benjamini–Hochberg correction (Benjamini and Hochberg 1995), a procedure for correcting false discovery rate with Type 1 errors. In the results that follow, we report mean Likert-scale ratings for each dialect in tables, and provide significance values in text.

When we consider the data in the aggregate sample, we find six statistically significant results for the main effect of dialect for the variables related to socioeconomic class (i.e., income, work, and family background). We found no significant effects for the main effect of label in the aggregate.

First we consider the data for the perception of profession. Participants rated the likelihood that each of the voices (Cuban, Colombian, Peninsular) in each of the family background conditions (Cuban, Colombian, Peninsular, no family label) worked (a) in a cell phone store, (b) at a coffee shop, (c) in a medical supplies company, (d) as a marketing executive, and (e) as an attorney. Because the results for “marketing executive” and “attorney” were almost numerically identical across conditions, we grouped them together for the statistical analysis, giving us a total of four professional categories. In the aggregate sample we found no significant interactions between dialect and label and no significant main effects of the label. However, we found statistically significant main effects for dialect for all professions except for “medical supplies company.” In the aggregate, we found that participants significantly favored the Peninsular voice for the white-collar professions and significantly disfavored the Peninsular voice for the non-white-collar professions. For example, the Peninsular voice was significantly less likely to be perceived as working in a coffee shop or as an employee in a cell phone store, compared with the Colombian and Cuban voices (p < 0.001). For the combined white-collar category, which included the perceptual ratings for “marketing executive” and “attorney,” the Peninsular voice was rated significantly higher than either of the other dialects (p = 0.007). There was no significant difference between the Peninsular, Cuban, and Colombian voice for the profession “medical sales company.” Table 3 presents the mean Likert-scale ratings in the aggregate sample for each voice, averaged across the four label conditions, for each of the four professions tested.
Table 3

Mean Likert-scale results for perception of “profession,” by dialect

Coffee shop







Cell phone store







Medical supplies manager







White collar (marketing executive + attorney)







Figure 1 depicts all the perceptual ratings for the combined white-collar professions. Although there were no significant interactions between the label and the dialect, we have included all conditions in the graph in order to show the range of perceptions given in the aggregate for the white-collar professions.
Fig. 1

Perception of white-collar professions, by dialect and background label

For the domain of family origin, we combined our five test statements into three categories for statistical testing, as follows: (a) family values hard work, (b) family provides opportunities to get ahead + family invested a lot of money in education, and (c) comes from a family that was pretty poor + comes from a family in which the prior generation had little choice for work. The results of statistical analysis in the aggregate sample resemble the results for profession in that we found a consistent significant split between the Peninsular voice and the Colombian and Cuban voices. For “family values hard work,” we found no significant difference among dialect groups (p = 0.968). For the combined “opportunities + education” category, we found participants in the aggregate to perceive of the Peninsular voice as significantly more likely to have come from a family that provided opportunities to get ahead and invested money in education than the Cuban and Colombian voices (p < 0.001). In the combined category having to do with family wealth and opportunities for employment, the Peninsular voice was rated significantly lower than the Cuban and Colombian voices (p < 0.001). Table 4 presents the mean Likert-scale ratings for the aggregate sample for the family origin test statements.
Table 4

Mean Likert-scale results for perception of “family background,” by dialect

Family values hard work







Opportunity + education







Prior generation was poor + few choices for work







In addition to perceptions about employment and family background, we were also interested in perceptions of annual income. Results for the question “What do you believe this person’s current income to be?” are provided in Fig. 2. Mean incomes are plotted in terms of tens of thousands of dollars on the y-axis, while the x-axis depicts all dialect/family-origin permutations, listed in order from highest earning to lowest earning.
Fig. 2

Perceived annual income, by dialect and background label

The top four highest-earning permutations are those anchored in the Peninsular variety, such that the average salary for the Peninsular voice with Colombian parents was $68,100, the Peninsular voice with Spanish parents was $62,900, the Peninsular voice with Cuban parents was $59,700 and the unlabeled Peninsular voice earned $58,500. The average salary for the Peninsular voice across family background permutations was $62,300. The average salary assigned to the Cuban voice was just $49,675, while the average salary assigned to the Colombian voice was $49,175. The main effect of language variety for annual income was significant (p < 0.001), with the Peninsular voice being assigned significantly more money than the Cuban and Colombian voices.

When we move from the aggregate sample to the results for the Latin@ identified participants only, we find twenty significant effects of the dialect, five significant effects of the label, and one significant interaction between dialect and label. We explain these findings in turn.

The results for the Latin@ participants’ perceptions of profession closely resemble the results of the aggregate for the professional traits. In the first statistical test, we compared ratings of the Peninsular variety with an average of the Cuban and Colombian ratings. Latin@ participants very strongly ranked the Peninsular voice below the other varieties for coffee shop (p < 0.001) and for cell phone store (p < 0.001) and very strongly ranked the Peninsular voice above the other varieties for the combined white-collar category (p < 0.001). When we consider the Cuban and Colombian voices separately in one-on-one comparisons with the Peninsular voice, we find similar results. When comparing the Peninsular voice with the Colombian voice, Latin@ participants found the Colombian voice to be significantly more likely to work in the coffee shop (p < 0.001) and cell phone store (p < 0.001) and found the Spaniard significantly more likely to be employed in the medical supplies company (p = 0.042) and in the white collar professions (p < 0.001). We found the exact same results with the same p-values when comparing the Cuban voice with the Peninsular voice, except there was no significant difference for medical supplies company. There were no significant differences between the Colombian and the Cuban voices for the professional traits when testing for the main effect of dialect for the Latin@ participants. Mean Likert-scale results for Latin@ participant responses for each voice in each professional category are provided in Table 5. The values for Spain are bolded in order to indicate significant differences from the combined Cuba/Colombia voice condition.
Table 5

Mean Likert-scale results for perception of “profession,” by dialect (Latino participants only)

Coffee shop







Cell phone store







Medical supplies manager







White collar (marketing executive + attorney)







Latin@ participants’ perception of dialect in terms of family background resembles the results observed in the aggregate sample. First, we found no significant differences for the main effect of dialect for the family origin trait “values hard work.” All three groups were rated highly for this trait. For the statistical test comparing the Peninsular voice with the combined Cuban/Colombian voice, we found that Latin@ participants rated the Spaniard higher than the others for the combined “opportunities + education” condition (p < 0.001) and the Spaniard significantly lower on the condition combining “family was poor” with “family had few choices for employment” (p < 0.001). When we analyze the three voices separately, we find the same results, with Latin@ participants rating the Peninsular voice significantly higher than the Colombian (p < 0.001) and the Cuban (p < 0.001) for the “opportunities + education” condition and rating the Peninsular voice significantly lower than Colombian (p < 0.001) and Cuban (p < 0.001) for the condition combining “was poor” and “few choices for employment.” For the latter condition, Latin@ participants also rated the Cuban voice significantly higher than the Colombian voice (p = 0.020). Mean Likert-scale results for these conditions are found in Table 6.
Table 6

Mean Likert-scale results for perception of “family background,” by dialect (Latino participants only)

Family values hard work







Opportunity + education







Prior generation was poor + few choices for work







For the combined family condition “was poor + few choices for employment” we also found three significant main effects for the background labels that were randomly assigned to the dialect voices. First, compared with the family background label “Spain,” the “Colombia” family background label had a significant raising effect (p = 0.021) for “came from a family that was poor” and “came from a family that had few opportunities for work.” The “Cuba” family background label also had a significant raising effect for this condition compared with the “Spain” background label (p = 0.004). We find the effect again when comparing the combined “Cuba/Colombia” family background labels with the “Spain” family background label (p = 0.005). In other words, when Latin@ participants believed that the voice they were hearing had parents from either Cuba or Colombia, they were significantly more likely to rate the speaker as coming from a “poor” family, compared with believing the speaker’s parents were from Spain.

In terms of annual income, Latin@ participants again favored the Peninsular voice, assigning it an average of $64,100 per year compared with $51,950 for the Cuban voice and $49,175 for the Colombian voice. Mean attributed incomes, averaged across the four background label conditions per dialect, are depicted in Fig. 3.
Fig. 3

Perceived annual income, by dialect

The income assigned to the Peninsular voice was significantly higher than that assigned to either the Cuban voice (p < 0.001) or the Colombian voice (p < 0.001). The Cuban and Colombian assigned salaries were not significantly different from each other (p = 0.241). For the annual incomes assigned by Latin@ participants, we found two significant main effects of the label. When participants made guesses about annual income for the voices they heard, believing the speaker’s parents came from Colombia had a significant raising effect compared with believing the speaker’s parents came from Cuba (p = 0.019). This is true even though the Cuban voice was assigned slightly higher incomes across conditions than the Colombian voice, though again this difference in assigned income was not significant. The second significant main effect of the label is between “Spain” and “Cuba.” Irrespective of the voice (Peninsular, Cuban, Colombian) the participants heard, the presence of the background information suggesting that the speaker’s parents were from Spain had a significant raising effect on income (p = 0.031). There was no significant difference between the family background labels “Spain” and “Colombia” (p = 0.913) for annual income.

Finally, our statistical analysis shows a significant interaction between dialect (Peninsular, Colombian, Cuban) and label (Spain, Colombia, Cuba, no label) for assigned annual income. First, the effect of the “Spain” and “Colombia” labels are more positive when applied to Cuban and Colombian voices than when applied to the Peninsular voice. Similarly, the effect of Colombian and Cuban labels was more positive when applied to the Peninsular voice than to the Colombian voice. The interaction shows that the effect of national-origin labels on income judgments varies across the three dialects. In other words, the effect of the label depends on the dialect with which it is paired. Thus, it is not the case that the family background label “Spain” drives income judgments up unilaterally, despite the strong main effect of the Peninsular dialect on perceived income and despite the positive effect of the “Spain” label as a main effect compared with the Cuban label.

Race, language ideology, and the meanings of Spanish in Miami

The findings obtained in this experimental study of linguistic perceptions show that Miami-Dade’s young Latin@s exhibit divergent perceptions of national origin varieties of Spanish and that they use sociolinguistic differences to make predictive judgments about nonlinguistic, social attributes related to socioeconomic class, including family wealth, personal income, and profession. Our statistical analyses of these judgments revealed three kinds of significant effects: (1) main effects of language variety, (2) main effects of the background label, and (3) interaction effects.

For the main effect of language variety, we found the Peninsular dialect to be strongly and consistently associated with family wealth, profession, and earning potential, both in the aggregate sample and among Latin@ judges. Latin@ listeners assigned significantly more money per year to the Peninsular voice, irrespective of the parents’ ostensible country of origin. They also strongly placed the voice in the white-collar professions of “attorney” and “marketing executive,” while strongly disassociating the Peninsular voice with family backgrounds that were economically disadvantaged. On the one hand, these findings are unsurprising in light of pervasive Eurocentric ideologies, colonialist ideologies that construct Spain as la madre patria, and the ideological workings of the Real Academia de la Lengua Española and other institutions that support “purist” language ideologies. And at first blush, it may also appear that participants in this study were simply making probabilistic guesses rooted in empirical realities; Spain is after all a middle-class country with all of the economic and political protections of the European Union. Cuba is after all a Caribbean nation rife with long-standing economic and political turmoil. But this reading does not account for the empirical reality in which Cuban Americans in Miami-Dade are both the numerical majority and the political elite. Sociologists studying Miami-Dade County (e.g., Stepick et al. 2003) have remarked on the unprecedented economic and political success of the Cuban American community in Miami-Dade. Moreover, national demographic statistics (Brown and Patten 2013) on the economic success of national origin Latin@ groups in the United States point to the success of Cuban Americans. For example, compared with the general US Latin@ population, Cuban Americans are more educated. Of Cubans over the age of twenty-five, 25% percent hold at least a bachelor’s degree, compared with 13% for all other US Latin@s. In terms of income, Cuban Americans over the age of sixteen earn a median annual salary of $24,400, which is less than the median income for the general US population ($29,000) but more than the median income for the general US Latin@ population ($20,000). And compared with the general US Latin@ population, Cuban Americans are more likely to have health insurance, more likely to own a home, and less likely to live in poverty. We appreciate that differences in racio-ethnic identity and immigration status across Latin@ national origin groups in the United States make these comparisons problematic at best. We make them nevertheless, in order to imagine alternative ways Cuban American language could be perceived but was not. Similarly, at the time our experiment was conducted, Spain was several years into its great recession (la crisis), which famously left about 50% of Spaniards under the age of twenty-five out of work (Eurostat 2015). Our point is not that the linguistic perceptions we found have no basis in empirical reality. Instead, we submit that ideologies about language, which are imbricated with ideologies about nation, class, ethnicity and other social formations, help construct the social difference that is in turn reflected as “real” in linguistic perceptions.

Though far less powerful perceptual cues than the voices themselves, the terms “Spain,” “Cuba,” and “Colombia,” also exerted a significant effect on perception when assumed to be countries of origin for the ostensible parents of the speakers heard in the experiment. These effects emerged when participants made predictive judgments related to socioeconomic class, including predictions about income and wealth of the speaker’s family of origin. Though we observed fewer—and less statistically powerful—effects for the label than for the dialect, we are surprised that this manipulation affected perception at all, considering how briefly participants saw the information about speaker’s family background, how much cognitive work was involved in completing the rest of the experiment, and how irrelevant this information actually is in making judgments about the voices. The significant interaction between the voice condition and the label condition for annual income suggests that listener perceptions are based on multiple sources of information, some found in what linguists call the “speech stream” or “acoustic signal,” and some in other nonlinguistic information known about the speaker.

Irvine (1989, p. 53) defines language ideology as “the cultural system of ideas about social and linguistic relationships, together with their loading of moral and political interests.” This definition is useful for our conversation because it suggests not only that language varieties are socioculturally figured in particular ways, but also that they are figured as a part of a broader system of meaning. That is, Peninsular Spanish signifies only in relation to Cuban Spanish and in relation to Colombian Spanish, and so on. Because language varieties act indexically for other social meanings, and because this indexical link is tended to so carefully by language academies, educational systems, and in everyday discourse and interaction, perceptions of individuals based on speech are practically overdetermined. We believe this is especially the case in Spanish-speaking Miami-Dade not only because of the presence of the county’s immense Spanish dialect diversity and the language ideologies from Spain and Latin America that attend it, but also because questions of language are already so fraught with signification in South Florida because of the presence of US-based language policies, politics, and ideologies that limit and problematize multilingualism while centralizing the assumption of English monolingualism.

In the introduction to this special edition, Mahler notes that the Anglo White hegemony we presume to be in place in most, if not all, regional settings in the United States is diminished in South Florida’s Miami-Dade County. Here, two points must be made in relation to language in general and to our findings in particular. First, although we agree that the widespread use of Spanish in Miami-Dade poses a challenge to the hegemonic Anglo White assumption of English monolingualism, we must pause to note the ways in which Anglo White linguistic hegemony shapes Miami-Dade’s language scene. The Miami-Dade Public School system offers a curriculum that is overwhelmingly English-only, despite the fact that the county was a pioneer in the bilingual education movement in the 1960s (Carter and Lynch 2015) and despite the fact that the vast majority of students who enter school learn Spanish in the home first. Therefore, Miami-born Latin@s receive the message that educational and sociocultural success is tied to English monolingualism. This fact is borne out in the literature on language shift in the county, which shows Spanish to be retreating generation by generation among the Miami-born (Porcel 2006; Portes and Schauffler 1996; Zurer Pearson and McGee 1993). Our point is that although the widespread use of Spanish in the region may give one the impression that Anglo White linguistic interests are not centralized in public life, Miami-Dade institutions are for the most part English-only. Thus, while Peninsular, Cuban, Colombian and other varieties of Spanish are in a dialectical relationship with each other, the Spanish language is itself in a dialectical relationship with English, living all the time under the specter of Anglo White linguistic, political, and sociocultural interests. That is, ideologies of English monolingualism (Macías 1985; Wiley and Lukes 1996; Santa Ana 2002) already problematize Spanish speaking in South Florida, creating more tension around questions of language and linguistic choice. We believe that it is against the backdrop of English dominance that language perceptions are so pronounced and so charged. Second, although Anglo White hegemony may be diminished in Miami-Dade, new forms of inequality articulated around intra-ethnic socioeconomic class differences and differences in national origin may be exacerbated, as our data suggest. Spanish is a potentially powerful unifying force in Miami-Dade, especially among immigrants, but language ideologies that demote certain varieties of Spanish along the lines of socioeconomic class and national origin limit this potential. It is incumbent on scholars of language in US Latin@ communities to understand how these inequalities operate in the context of sociolinguistic variation and to educate the public about the realities of language and language variation.

In terms of the ways in which the category “Latin@” is being constructed in Miami-Dade and beyond, questions of language are not superficial or ancillary, but fundamental. In Miami-Dade, Spanish is an undeniably vital part of the kinds of latinidad taking shape here. But just as the category “Latin@” cannot be taken to be homogeneous, nor can the languages spoken by the people who inhabit that category. In Miami-Dade then, the question is not “Do you speak Spanish?” but instead “What kind of Spanish do you speak?” The data we have presented here indicate that the answer to this question carries important implications for the ways people perceive of one another. We thus encourage scholars studying the processes of racial formation (e.g., Omi and Winant 1994) related to the racio-ethnic category “Latin@” described in the introduction to this volume to consider the ways language shapes its conditions of possibility. Here we echo Zentella’s (1995, p. 13) call for an “anthropolitical linguistics” in which linguists and anthropologists are committed to combating the negative and disaffirming language ideologies that affect US Latin@s. We also acknowledge that the work of disarticulating and dislodging damaging language ideologies requires the coordinated efforts of a broad coalition of scholars, not only linguists and anthropologists. To this end, we hope that language figures as a central and primary site of inquiry for interdisciplinary scholarship on Latin@ communities and identities in the United States.


  1. 1.

    Although these turns of phrase suggest that these linguistic features are socially stigmatized or are inherently wrong or incorrect, linguists recognize that the phonological process of lenition or consonant “weakening” is a natural process in language change. Furthermore, similar patterns are not condemned, for example, in Rioplatense varieties of Spanish.

  2. 2.

    For participants who identified as Hispanic/Latino, we collected information on family national origin background. Although there were not enough participants in each group to run separate statistical analyses, we report the national-origin background information here: 91 (43%) Cuban, 29 (14%) Central American, 2 (10%) Colombian, 9 (4%) Dominican, 4 (2%) Mexican, 14 (7%) Puerto Rican, 13 (6%) Venezuelan, 20 (10%) South American (other than Venezuela or Colombian), 8 (4%) Spanish (Spain).

  3. 3.

    In this article we use the term Anglo White to refer to the population of typically monolingual English-speaking people of European descent who are variably referred to in popular discourse as “Caucasian,” “Anglo,” and “White.” We do not use the term “White,” which is commonly used in academic writing on ethnicity, since many people in Miami who inhabit the racio-ethnic category “Latino” consider themselves racially “White.” We also avoid the term “Anglo,” which tracks in some parts of the US as a euphemism for “White.”

  4. 4.

    Participants did not hear every condition for two reasons. First, the number of conditions tested made it logistically impossible to test every condition with every participant for reasons related to participant attrition. Second, we limited the number of conditions heard by each participant in order to increase the plausibility of the manipulation. That is, the more mismatched dialect/label permutations a speaker hears, the more likely they are to catch on to the experimental manipulation.



We are grateful to the three anonymous external reviewers for their invaluable feedback on the manuscript, to Andrew Lynch and Melissa Baralt for their feedback at various stages of the research and writing process, and to David Neal for his expertise in study design and statistical analysis. We are especially indebted to Scott Schwenter for suggesting that we pursue this line of research, and to Sarah Mahler for including our study in this special edition and for providing invaluable feedback on several drafts of the manuscript. Any errors herein are our own.


  1. Alfaraz, G. 2002. Miami Cuban Perceptions of Varieties of Spanish. In Handbook of Perceptual Dialectology, vol. 2, ed. D. Long, and D. Preston, 1–12. Amsterdam: John Benjamins.Google Scholar
  2. Alfaraz, G. 2014. Dialect Perceptions in Real Time: A Restudy of Miami Cuban Perceptions. Journal of Linguistic Geography 2 (2): 74–86.CrossRefGoogle Scholar
  3. Anderson, B. 1983. Imagined Communities: Reflections on the Origin and Spread of Nationalism. New York: Verso Books.Google Scholar
  4. Babel, A. 2014. Stereotypes Versus Experience: Indexing Regional Identity in Bolivian Valley Spanish. Journal of Sociolinguistics 18 (5): 604–633.CrossRefGoogle Scholar
  5. Baugh, J. 2003. Linguistic Profiling. In Black Linguistics: Language, Society, and Politics in Africa and the Americas, ed. S. Makoni, G. Smitherman, A.F. Ball, and A.K. Spears, 155–168. London: Routledge.Google Scholar
  6. Benjamini, Y., and Y. Hochberg. 1995. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society (Series B) 57: 289–300.Google Scholar
  7. Blom, J.P., and J.J. Gumperz. 1972. Social Meaning in Linguistic Structures: Code Switching in Northern Norway. In Directions in Sociolinguistics, ed. J.J. Gumperz, and D. Hymes, 407–434. New York: Holt, Rinehart, and Winston.Google Scholar
  8. Boersma, P., and D. Weenink. 2013. Praat: Doing Phonetics by Computer (Computer program).Version 5.3.23.
  9. Bourdieu, P. 1984. Distinction: A Social Critique of the Judgement of Taste. Cambridge, MA: Harvard University Press.Google Scholar
  10. Bourdieu, P. 1991. Language and Symbolic Power. Cambridge, MA: Harvard University Press.Google Scholar
  11. Bradac, J.J., and R. Wisegarver. 1984. Ascribed Status, Lexical Diversity, and Accent: Determinants of Perceived Status, Solidarity, and Control Of Speech. Journal of Language and Social Psychology 3 (4): 239–255.CrossRefGoogle Scholar
  12. Brown, A., and M.H. Lopez. 2013. Mapping the Latino Population by State, County, and City. Washington, DC: Pew Research Center.Google Scholar
  13. Brown, A., and E. Patten. 2013. Hispanics of Cuban Origin in the United States, 2011. Washington, DC: Pew Research Center.Google Scholar
  14. Bucholtz, M., N. Bermudez, L. Edwards, V. Fung, and R. Vargas. 2007. Hella Nor Cal or Totally So Cal?: The Perceptual Dialectology of California. Journal of English Linguistics 35 (4): 325–352.CrossRefGoogle Scholar
  15. Carter, P.M. 2013. Shared Spaces, Shared Structures: Latino Social Formation and African American English in the U.S. South. Journal of Sociolinguistics 17 (1): 66–92.CrossRefGoogle Scholar
  16. Carter, P.M., and A. Lynch. 2014. Ideological Dimensions of Speech Perception in Bilingual Miami. Madison, WI: Paper delivered at Workshop on Spanish Sociolinguistics. April.Google Scholar
  17. Carter, P.M., and A. Lynch. 2015. Multilingual Miami: Current Trends in Sociolinguistic Research. Language and Linguistics Compass 9 (9): 369–385.CrossRefGoogle Scholar
  18. Cedergren, H. 1973. The Interplay of Social and Linguistic Factors in Panama. PhD dissertation, Cornell University, Ithaca, NY.Google Scholar
  19. Docherty, G., C. Langstrof, and P. Folkes. 2013. Listener Evaluation of Sociophonetic Variability: Probing Constraints and Capabilities. Linguistics 51 (2): 355–380.CrossRefGoogle Scholar
  20. Duany, J. 1998. Reconstructing Racial Identity: Ethnicity, Color, and Class among Dominicans in the United States and Puerto Rico. Latin American Perspectives 25 (3): 147–172.CrossRefGoogle Scholar
  21. Eckert, P. 2000. Linguistic Variation as Social Practice. Malden, MA: Blackwell.Google Scholar
  22. Eckert, P. 2005. Variation, Convention, and Social Meaning. Annual Meeting of the Linguistic Society of America. Vol. 7. Oakland, CA: Linguistic Society of America.Google Scholar
  23. Escobar, A.M., and K. Potowski. 2015. El Español de los Estados Unidos. Cambridge, UK: Cambridge University Press.Google Scholar
  24. Eurostat. 2015. Youth Unemployment. Eurostat: Statistics Explained. 1 September.
  25. Feagin, C. 1979. Variation and Change in Alabama English: A Sociolinguistic Study of the White Community. Washington, DC: Georgetown University Press.Google Scholar
  26. Fought, C. 2002. California Students’ Perceptions of, You Know, Regions and Dialects. In Handbook of Perceptual Dialectology, vol. 2, ed. D. Long, and D. Preston, 113–134. Amsterdam: John Benjamins.Google Scholar
  27. Fridland, V., and K. Bartlett. 2006. Correctness, Pleasantness, and Degree of Difference Ratings across Regions. American Speech 81 (4): 358–386.CrossRefGoogle Scholar
  28. Fridland, V., K. Bartlett, and R. Kreuz. 2004. Do You Hear What I Hear? Experimental Measurement of the Perceptual Salience of Acoustically Manipulated Vowel Variants by Southern Speakers in Memphis. TN. Language Variation and Change 16 (1): 1–16.Google Scholar
  29. Fuller, J. 2012. Bilingual Pre-teens: Competing Ideologies and Multiple Identities in the US and Germany. London: Routledge.Google Scholar
  30. García, O., I. Evangelista, M. Martínez, C. Disla, and B. Paulino. 1988. Spanish Language Use and Attitudes: A Study of Two New York City Communities. Language in Society 17 (4): 475–511.CrossRefGoogle Scholar
  31. Guy, G. 2005. Grammar and Usage: A Variationist Response. Language 81 (3): 561–563.CrossRefGoogle Scholar
  32. Horvath, B. 1985. Variation in Australian English: The Sociolects of Sydney. Cambridge Studies in Linguistics London 45: 1–200.Google Scholar
  33. Irvine, J. 1989. When Talk Isn’t Cheap: Language and Political Economy. American Ethnologist 16 (2): 248–267.CrossRefGoogle Scholar
  34. Labov, W. 1966. The Social Stratification of English in New York City. Washington DC: Center for Applied Linguistics.Google Scholar
  35. Lambert, W.E., R.C. Hodgson, R.C. Gardner, and S. Fillenbaum. 1960. Evaluational Reactions to Spoken Languages. Journal of Abnormal and Social Psychology 60 (1): 44–51.CrossRefGoogle Scholar
  36. Lavandera, B. 1978. Where Does the Sociolinguistic Variable Stop? Language in Society 7 (2): 171–182.CrossRefGoogle Scholar
  37. Lippi-Green, R. 1997. English with an Accent: Language, Ideology, and Discrimination in the United States. London: Routledge.Google Scholar
  38. Lopez, M. H., A. Gonzalez-Barrera, and D. Cuddington. 2013. Diverse Origins: The Nation’s 14 Largest Hispanic-Origin Groups. Washington, DC: Pew Research Center.
  39. Lynch, A. 2009. A Sociolinguistic Analysis of Final/s/ in Miami Cuban Spanish. Language Sciences 31 (6): 766–790.CrossRefGoogle Scholar
  40. Macías, R. 1985. Language and Ideology in the United States. Social Education 49 (2): 97–100.Google Scholar
  41. Massey, D., and G. Lundy. 2001. Use of Black English and Racial Discrimination in Urban Housing Markets: New Methods and Findings. Urban Affairs Review 36 (4): 470–496.CrossRefGoogle Scholar
  42. Mendoza-Denton, N. 2008. Homegirls: Language and Cultural Practice among Latina Youth Gangs. Malden, MA: Blackwell.CrossRefGoogle Scholar
  43. Modaressi, Y. 1978. A Sociolinguistic Investigation of Modern Persian. PhD dissertation, Lawrence, KS: University of Kansas.Google Scholar
  44. Mojica de León, C.M. 2014. Una mirada hacia las actitudes lingüística en Puerto Rico. Bergen Language and Linguistics Studies 5: 1249–1315.CrossRefGoogle Scholar
  45. Naro, A. 1981. The Social and Structural Dimensions of a Syntactic Change. Language 57 (1): 63–98.CrossRefGoogle Scholar
  46. Niedzielski, N. 1999. The Effect of Social Information on the Perception of Sociolinguistic Variables. Journal of Language and Social Psychology 18 (1): 62–85.CrossRefGoogle Scholar
  47. Niedzielski, N., and Preston. 2000. Folk Linguistics. In Trends in Linguistics Studies and Monographs, vol. 122, ed. W. Winter. Berlin: Mouton de Gruyter.Google Scholar
  48. Omi, M., and H. Winant. 1994. Racial Formation in the United States: From the 1960s to the 1990s, 2nd ed. London: Routledge.Google Scholar
  49. Otheguy, R., O. García, and A. Roca. 2000. Speaking in Cuban: The Language of Cuban Americans. In New Immigrants in the United States: Readings for Second Language Educators, ed. S. McKay, and S.C. Wong, 165–188. Cambridge, UK: Cambridge University Press.Google Scholar
  50. Otheguy, R., A.C. Zentella, and D. Livert. 2007. Language and Dialect Contact in Spanish in New York: Toward the Formation of a Speech Community. Language 83 (4): 770–802.CrossRefGoogle Scholar
  51. Otheguy, R., and A.C. Zentella. 2012. Spanish in New York: Language Contact, Dialect Leveling, and Structural Continuity. Oxford, UK: Oxford University Press.CrossRefGoogle Scholar
  52. Poplack, S. 1980. Sometimes I’ll Start a Sentence in Spanish y Termino en Español: Toward a Typology of Code-switching. Linguistics 18 (7–8): 581–618.Google Scholar
  53. Porcel, J. 2006. The Paradox of Spanish among Miami Cubans. Journal of Sociolinguistics 10 (1): 93–110.CrossRefGoogle Scholar
  54. Portes, A., and R. Schauffler. 1996. Language and the Second Generation: Bilingualism Yesterday and Today. In The New Second Generation, ed. A. Portes, 8–29. New York: Russell Sage.Google Scholar
  55. Preston, D.R. 1989. Perceptual Dialectology. Dordrecht: Foris.CrossRefGoogle Scholar
  56. Preston, D.R. 1993. Folk Dialectology. In American Dialect Research, ed. D.R. Preston, 333–377. Amsterdam: John Benjamins.CrossRefGoogle Scholar
  57. Preston, D.R. 1996. Whaddayaknow? The Modes of Folk Linguistic Awareness. Language Awareness 5 (1): 40–74.CrossRefGoogle Scholar
  58. Preston, D. R. 1999. A Language Attitude Approach to the Perception of Regional Variety. In Handbook of Perceptual Dialectology, ed. D. R. Preston, vol. 1. Trans. B. E. Evans. Amsterdam: John Benjamins.Google Scholar
  59. Preston, 2002. Perceptual Dialectology: Aims, Methods, Findings. In Trends in Linguistics Studies and Monographs, vol. 137, ed. W. Winter, and W. Bisand, 57–104. Berlin: Mouton de Gruyter.Google Scholar
  60. Purnell, T., W. Idsari, and J. Baugh. 1999. Perceptual and Phonetic Experiments on American English Dialect Identification. Journal of Social Psychology 18 (1): 10–30.Google Scholar
  61. Quesada Pacheco, M.A. 2015. División dialectal del Español de América según sus hablantes: Análisis dialectológico perceptual. Boletín de Filología 49 (2): 257–309.CrossRefGoogle Scholar
  62. Rickford, J.R., and S. King. 2016. Language and Linguistics on Trial: Hearing Rachel Jeantel (and Other Vernacular Speakers) in the Courtroom and Beyond. Language 92 (4): 948–988.CrossRefGoogle Scholar
  63. Santa Ana, O. 2002. Brown Tide Rising: Metaphors of Latinos in Contemporary American Public Discourse. Austin: University of Texas Press.Google Scholar
  64. Shuy, R.W., W. Wolfram, and W.K. Riley. 1968. Field Techniques in an Urban Language Study, vol. 3. Washington DC: Center for Applied Linguistics.Google Scholar
  65. Silverstein, M. 2003. Indexical Order and the Dialectics of Sociolinguistic Life. Language & Communication 23 (3): 193–229.CrossRefGoogle Scholar
  66. Slomanson, P., and M. Newman. 2004. Peer Group Identification and Variation in New York Latino English Laterals. English World Wide 25 (2): 199–216.CrossRefGoogle Scholar
  67. Sobrino Triana, R., L.E. Montero Bernal, and A.J. Menéndez Pryce. 2014. Actitudes lingüísticas en Cuba. Cambios positivos hacia la variante nacional de lengua. Bergen Language and Linguistics Studies 5: 290–408.CrossRefGoogle Scholar
  68. Stepick, A., G. Grenier, M. Castro, and M. Dunn. 2003. This Land is Our Land: Immigrants and Power in Miami. Berkeley: University of California Press.Google Scholar
  69. Suárez Büdenbender, E.M. 2013. “Te conozco bacalao”: Investigating the Influence of Social Stereotypes on Linguistic Attitudes. Hispania 96 (1): 110–134.CrossRefGoogle Scholar
  70. Tamasi, S. 2003. Cognitive Patterns of Linguistic Perceptions. PhD dissertation, Athens: University of Georgia.Google Scholar
  71. Tetel Andresen, J., and P.M. Carter. 2016. Languages of the World: How History, Culture, and Politics Shape Language. Malden, MA: Wiley-Blackwell.Google Scholar
  72. Trudgill, P. 1974. The Social Differentiation of English in Norwich. Cambridge, MA: Cambridge University Press.Google Scholar
  73. US Census Bureau. 2010. 2010 Census.
  74. Veltman, C. 1988. Modeling the Language Shift Process of Hispanic Immigrants. International Migration Review 22 (4): 545–562.CrossRefGoogle Scholar
  75. Wiley, T., and M. Lukes. 1996. English-Only and Standard English Ideologies in the U.S. TESOL Quarterly 30 (3): 511–535.CrossRefGoogle Scholar
  76. Wolfram, W., P.M. Carter, and B. Moriello. 2004. Emerging Hispanic English: New Dialect Formation in the American South. Journal of Sociolinguistics 8 (3): 339–358.CrossRefGoogle Scholar
  77. Wolfram, W., and N. Schilling. 2015. American English: Dialects and Variation, 3rd ed. Malden, MA: Wiley-Blackwell.Google Scholar
  78. Zentella, A.C. 1995. The “Chiquitafication” of U.S. Latinos and Their Languages, or: Why We Need an Anthropolitical Linguistics. Proceedings of Symposium about Language and Society III. Austin: University of Texas Press.Google Scholar
  79. Zentella, A.C. 1997. Growing Up Bilingual: Puerto Rican Children in New York. Malden, MA: Wiley Blackwell.Google Scholar
  80. Zentella, A.C. 2014. TWB (Talking while Bilingual): Linguistic Profiling of Latina/os, and Other Linguistic Torquemadas. Latino Studies 12 (4): 620–635.CrossRefGoogle Scholar
  81. Zurer Pearson, B., and A. McGee. 1993. Language Choice in Hispanic-background Junior High School Students in Miami: A 1988 Update. In Spanish in the United States: Linguistic Contact and Diversity, ed. A. Roca, and J. Lipski. Berlin: Mouton de Gruyter.Google Scholar

Copyright information

© Macmillan Publishers Ltd., part of Springer Nature 2018

Authors and Affiliations

  1. 1.Florida International UniversityMiamiUSA
  2. 2.University of Texas at AustinAustinUSA

Personalised recommendations